66 



SCIEXCE- GOSSIP. 



intr a correct representation of the lights and 

 shades of our object. 



In questions of colour it is always as well to 

 deal first of all with colours in their gre; I 

 purity — that is. white light decomposed int its 

 constituents. In fig. 1 we have a drawing of the 

 prismatic spectrum, and in fig. 2 I hare represe: 



v a line the apparent brightness or luminosity to 

 t he human eye of the various parts of the spectrum. 

 this line indicating by its height above the ' i - - 

 line the varying depths of light and shade. A 

 perfect photograph of the spectrum should then 

 give us a result something like this, but fig. 3, 

 which represents a photograph of the spectrum 

 taken upon an ordinary plate, shows us how far 

 away from truth is really the ordinary photograph. 

 We here see that nearly all the photographic 

 action has taken place in the blue and invisible 

 violet, whilst the highest visual luminosity of 

 the spectrum — i.e. in the yellow — is represented as 

 black. The difference shown is so great that at 

 first we wonder how ordinary photography can 

 give any representation worthy of the na m e. 

 Were it not for the considerable amount of white 

 light reflected by nearly all coloured objects in 

 nature, the use of photography for pictorial repre- 

 sentation would be practically impossible. Never- 

 theless, this want of orthochromatism or luminosity 

 sensitiveness is serious in many ways. For in- 

 stance, the ordinary plate does not distinguish the 

 luminous value of blue sky from that of white 

 cloud, but it is more especially when photographing 

 close objects, notably where the colours are 

 brilliant, that the deficiency of the ordinary plate 

 is most felt. 



:lear, then, from these facts, that what we 

 have to do is to secure such photographic action 

 upon our plates as will give us densities in our 

 negatives directly proportionate to the visual in- 

 tensities of the bight reflected from the object, 

 ignoring totally, no matter what they may be, the 

 colours of the object. 



Xow. nearly all photographic emulsions are 

 sensitive ro some extent to all the visible spectrum. 

 but such sensitiveness differs very widely from the 

 eye-luminosity of the curve shown above. 



There are two methods by which we might 

 hope to improve our results. First, by so altering 

 the constitution of the emulsion as to render it 

 more sensitive to green and red ; and. secondly, by 

 filtering or screening down the action of the blue 

 and violet so as to allow the red and green sensi- 

 tiveness of the emuLsion to impress the plate in its 

 proper proportion. 



With regard to the first method, a very great 

 deal has been done by various workers to improve 

 the sensitive film. In 1873. Professor Yogel dis- 

 covered that certain dyes incorporated with the 

 emulsion had the property of sensitising the result- 

 ing film to the portions of the spectrum absorbed 

 by the dye used. Not all dyes, however, pc— ss 



this property of increasing sensitiveness, and much 

 laborious experiment was necessary before plates 

 of commercial value were obtainable. In fig. 4 

 we have the sensitiveness of several commercial 

 brands of plates to the white light of the spectrum 

 represented by curves. The best of these curves — 

 that is. the one nearest approaching the curve of 

 eye-luminosities of the spectrum — is the last shown, 

 namely, that of the Cadett Lightning Spectrum 

 plate. On comparing this, however, with the ideal 

 curve, we find it is still far from perfect, the 

 blue and violet still being responsible for by far the 

 greater portion of the photographic action ; but 

 one will notice that now we have considerable sen- 

 sitiveness in the green and the red. and by cutting 

 down the blue and violet by a filter, it is possible to 

 secure, with a reasonable length of time of expo- 

 sure, an accurate representation of the spectrum. 

 Obviously the construction of a light filter which 

 will cut down the action of the green, blue, and 

 violet in exactly the right proportion is a very 

 difficult matter. The few early experimenters who 

 succeeded in doing this found it a very long and 

 wearisome task : for each trial of a filtering me- 

 dium negatives of the spectrum had to be taken 

 upon the sensitive plate, and then these negatives 

 had to be measured by a photometer, for the 

 amount of light passed. Only a very narrow band 

 of the spectrum could be measured at one time, 

 and twenty or thirty measurements of the different 

 parts of the spectrum were necessary, before we 

 could tell what amount of success had been ob- 

 tained. Such a method, therefore, was of little 

 use from the commercial point of view, on account 

 of the cost of the necessary skilled labour to con- 

 duct these delicate measurements. 



"illiam Abney has. however, by the inven- 

 tion of his Colour-Sensitometer, enabled us to pro- 

 duce light filters of great accuracy at so reasonable 

 a cost that every photographer can avail himself 

 of their use. It is a well-known fact in photo- 

 metry that, although the eye is a very bad judge of 

 the difference in intensity between two unequal 

 lights, or of the amount of fight transmitted by 

 two unequal densities in a negative, it is an excel- 

 lent judge of the equality of two lights, or of the 

 amount of light transmitted by two equal or nearly 

 equal densities on a photographic plate. Sir 

 William Abney availed himself of this fact in the 

 construction of his sensitometer. The principle 

 of the sensitometer is as follows : — Suppose we 

 take a white, a red, a yellow, a green, a blue- 

 green and a blue glass, and mount these in a row 

 in such a manner that we can photograph them 

 through the light filter we are adjusting, we shall 

 have, provided the glasses are suitably chosen, a 

 very fair representation of the spectrum ; with the 

 additional advantage that they will transmit 

 colours which are to some extent mixed colours, 

 and thereby the excess of one spectrum colour 

 mav be balanced bv the defect of another. If we 



