lO 



NA TURE 



[November 3, 1892 



How different is this from Clerk Maxwell's top. Clerk Max- 

 well selected for his top the purest of paper and pigments. He 

 endeavoured 10 match the spectral colours (considerably diluted). 

 He selected a scarlet red with a tinge of orange like orange-red 

 vermilion, lying in the spectrum one-third the way towards D, 

 between the lines C and D. His green was one- fourth the dis- 

 tance from E, between E and F, and resembled emerald green. 

 He also selected a blue violet midway between F and G, 

 which was imitated by that purest of colours — ultramarine. Now 

 let us try the given experiment under the favourable conditions 

 guaranteed by Maxwell's discs, viz., the purest of colours 

 painted on Whatman's paper. Taking up a disc of ultramarine 

 and another of pale (not orange) chrome yellow, and conceal- 

 ing half of one disc behind the other, on rotating the compound 

 disc so that the eye shall receive simultaneously blue and yellow 

 light, the result is not white or even practically white, but a 

 grey, tingid with yelloiv. By a careful adjustment, hiding more 

 of the yellow and exposing more of the blue (thereby altering 

 the proportions of the text), it is possible to get rid of this 

 yellowness and to obtain an absolutely neutral grey which it 

 might be possible to persuade some grown-up people repre- 

 sented white, but which on analysis yields 7 1^ per cent, black 

 to 284 per cent, white. This may be proved by revolving a 

 disc of black and white sectors in the above proportions, the 

 res\ilts in each case being identical. But even this result, un- 

 satisfactory as it is, does not apply to the passage quoted in 

 the text, in which no special conditions are observed. I main- 

 tain what is easily proved by experiment in less time than it takes 

 to write it, that when ordinary colours, e.g., gamboge and 

 Prussian blue, are used, the residual light is green. ^ 



I fear that already this letter i.s too long, and since I do 

 not wish to monopolize the space kindly placed at the dis- 

 posal of your correspondents, I must defer the consideration 

 of the annotations on soap films. The other points are dealt 

 with in the preface. Amy Johnson. 



52 Lower Sloane Street, S.W., October 12 



I DO not think that the observations on my review of " Sun- 

 shine " Require more than a very short answer. 



I considered that the authoress had not by any means cleared 

 the confusion which usually exists as to the meaning of the 

 expression "mixing of colours." It is applied both to the case 

 where two or more colours are seen superposed, e.g. by spin- 

 ning coloured paper where the resultant tint is due to the sum 

 of the separate colours in the constituents, and to the case of 

 mixed pigments where the resultant tint is that which is com- 

 mon to the constituents. Now as the common " paint box " 

 rule says that blue and yellow make green, that is that blue 

 and yellow pigments mixed produce a green pigment, it seems 

 tome very misleading tosay " Cover half (of your top) blue 

 and half yellow and you will see green." Of course it may 

 happen that the slight departure from white which will be ob- 

 served may be in a greenish direction, but it may also be inclined 

 towards pink, or, for anything I know, towards any other 

 colour. The one thing it will not do, however, is to make 

 a green such as is obtained by mixing the pigments, and such 

 as I fancy from the context any one would expect. C. V. B. 



The Photography of an Image by Reflection. 



The great utility of spark photography for obtaining .time 

 records of quickly-moving objects must be apparent to all 

 who know the experiments of Mr. C. Bell, Prof. Boys, and 

 Lord Rayleigh. By means of spark photography. the shadow of 

 any object such as a jet of water, a flying bullet, or a broken soap 

 film can be produced with perfect definition. The shadow of 

 the moving object illuminated by an electric spark is thrown on 

 to a sensitive plate in a dark room, and the plate is developed 

 in the usual manner. The process of spark shadow photography 

 will be found, I believe, of great service in physiological research. 

 With a view to try this I attached a long sensitive plate to the 

 traversing carriage of a chronograph ; the moving carriage closed 

 and opened the primary circuit of an induction coil at pre- 



' The purport of the experiment will be best understood if I state that it 

 rollows a series of chapters on colour, viz. : the rainbow, the spectrum, its 

 ecomposition by refraction and by reflection ; while the last chapter dis- 

 cusses and explains, with experiments, the question of spectral lights versus 

 pigments. The common surface papers, which the children are daily in the 

 habit of using, are then analysed by the prism, and found to be anything but 

 monochromatic. 



NO. T20T, VOL. 47] 



arranged equal intervals of time. In front of the moving plate a 

 frog's heart was placed in a slit on a screen ; at each break a 

 shadow of the heart was thrown on to the plate by means of the 

 induced spark. By this means thirty positions of the heart were 

 registered ; the pictures were all sharp and clear. I have also 

 used the same method for photographing the movements of in- 

 sects. 



Since these experiments which I showed during the University 

 Extension Meeting in Oxford this year, I have made several 

 attempts to get spark photographs of the front view of objects 

 (not their shadows). In my first experiments the objects were 

 illuminated by an electric spark, the image being received on a 

 plate in an ordinary camera. I found that so much useful light 

 was shut off by the lenses that only a dim picture could be 

 produced. A quartz lens was next tried and the results were 

 rather better. I then determined to use nolens, but in its place 

 a silvered mirror. A concave reflector madeby silvering a con- 

 cave lens of about 10 cm. diameter was so placed that it reflected 

 the image of a white paper star 7 cm. diameter, revolving about 

 60 times in a second, on to an ordinai-y photographic plate, the 

 total length traversed by the light being 80 cm. The star was 

 illuminated with a spark exactly similar to that used in the 

 previous experiment ; on development a good picture of the star 

 came out. The reflector was neitherwell made nor well silvered. 

 The idea was suggested by observing some spark photographs I 

 obtained of waves on the surface of mercury reflecting light. 

 When a steady light is used a photograph of any object is 

 readily obtained by reflection from a suitable mirror. Prob- 

 ably a steel surface m ould be best. The mirror and plate 

 were placed in a long box provided with a hole at one 

 end through which the light reflected from the object passed. 

 A few experiments made on living objects to test the time of 

 exposuie in Reflection Photography showed that in order to 

 avoid over-exposure, a very rapid shutter must be used. 



Frederick J. Smith. 



Trinity College, Oxford, October 25. 



Induction and Deduction. 



As your correspondent invites discussion on this subject I hope 

 you will allow me to repeat in anew form the views I expressed 

 upon it in your columns some months ago. I quite agree with 

 Mr. Russel in maintaining that " true induction is utterly un- 

 able to yield us any conclusion that is more than probable'and 

 approximate," understanding by induction inference from one 

 Or more special cases to a more general rule. But on the other 

 hand it appears to me that Miss Jones's criticism is quite de- 

 structive of Mr. Russel's interpretation of geometrical reasoning. 

 The point which both have missed I believe to be this, that a 

 proposition stated in given words, such as "the enunciation of 

 Euclid's pons asinorum dots not always and to every one con- 

 vey the same information ; and if it is meant in one sense its 

 digree of reliability, and the method by which it must be 

 proved, will be quite different from what they would be if it 

 were meant in another. There are at least three different kinds 

 of interpretation w hich may thus be put upon the proposition. 

 It may mean (i) tlie triangle used to illustrate this proposition 

 has equal sides ; therefore it has equal angles ; or (2) I have con- 

 ceived a triangle which has equal sides, therefore I have con- 

 ceived one which has equal angles ; or (3) the connotation ascribed 

 by the adjective "isosceles" implies the connotation "having 

 equal sides." 



It is not necessary for me here to dwell upon the distinction 

 between the first two. interpretations ; but the difference be- 

 tween either of them and the third is that this latter" gives us no 

 information about any real thing or concept, but only about 

 what is implied by using certain terms. And this latter kind of 

 information clearly does not require to be based upon any real 

 knowledge of things, but may be based solely on definitions of 

 words. Arguments with propositions interpreted only in 

 this sense are what I call symbolic arguments ; and symbolic 

 conclusions therefore give no real information unless they can 

 be interpreted by the aid of real assertions, such as " I can 

 conceive," or " There actually exist, things possessing the 

 ■connotations ascribed to these terms by their definitions." 



If this distinction has not before been recognized, it is because 

 in most logical discussions we can in this way give a real mean- 

 ing to our arguments. In elementary geometry, for example, 

 we can — with more or less effort — conceive things, or even 

 actually draw them, which answer to our definitions with 

 sufficient accuracy. And, indeed, the reason why "Euclid" 



