Contributions to the Study " ssed into good contact with the aluminium discs , served as a ms of transmitting current through the electrolyte between the js .
The frequency was 73 , the temperature 18 ' C. , and square root mean square value of current about 1 ampere .
No perceptible ierencewas observed in phase difference between potential and rent when the discs were rotated and at rest in the electrolyte .
'he conclusion is that the effect investigated in this paper takes e to develop , and is not fully developed with alternate currents of juencies sixteen and ninety-eight complete periods per second .
It i be increased by increasing the current density for a given film , and greatly influenced by temperature .
The metal aluminium with its ni is suitable for use as the plates of condensers , if due regard be en to current density and temperature .
It might in some cases found useful as an equivalent to a metallic resistance .
Vlessrs .
Simpson , Greenbank , and Davy , Student Demonstrators in t Siemens Laboratory , King 's College , London , have given me uable assistance in the experimental part , and in the working out results .
To these gentlemen I tender my thanks .
Contributions to the Study of ' Flicker .
' " By T. C. Porter , Eton College .
Communicated by Lord Rayleigh , F.R.S. Received May 13 , \#151 ; Read May 26 , 1898 .
Much work has already been done on this subject , though little of a antitative character .
Many observers have described the curious our sensations which rapid alternations of light and darkness can 3ite under certain conditions , admirably exemplified in the " spec-im tops .
" Foremost amongst those whose experiments and writings ve led to the present very general interest in the subjects of flicker , d of the sensation of light and colour , may be mentioned Helmholtz , ivanus Thompson , Shelford Bidwell , Henry , Charpentier , and tod ; whilst the first to try experiments on the relative sensitive-ss of the eye to flicker in light of different colours , seems to have en J. Plateau , who , however , employed pigments , and not the lours of the spectrum .
The writer 's first experiments were made to ascertain the exact lative rotations at which the flickei ' just vanishes in the different lours of the same spectrum , and were carried out ( a ) as suggested f Professor Rood in his 'Modern Chromatics , ' with a balanced , ackened , opaque disc , having a broad semicircular arc removed , 1(1 ( J\gt ; ) on a cardboard disc , half black , half white , viewed in the fferent colours of the spectrum of the second order of a Rowland s an diffraction grating of 14,434 lines to the inch .
Two sources ought were employed , ( a ) direct sunshine , ( \#163 ; \gt ; ) lime-light .
Ihe lesults Mr. T. C. Porter .
are conveniently expressed by curves ( fig. 1 ) .
along one axis of wcj , the spectrum is plotted , on the other the rate of rotation of the iSP \f S\amp ; 1/ T Solar \D Spectrum\c \3 Ihis last was found from the pitch of the note , obtained by blowing1 gently through a known number of equidistant holes pierced in ' circumference of the disc , and comparing these notes with those c1 set of standard forks .
Where the lime-light was used , great pH ' were taken to keep the intensity of the light constant through f each experiment , though in these first experiments the intens varied for the different curves .
It may be well to point out tl ' owing to the nature of the spectrum of lime , the orange is somewl brighter and the blue and violet darker , relatively , than they are sunshine .
It is for a similar reason , as well as for others too obvio-to mention , that the arc light was not employed .
It will be set that in fig. 1 , the intervals of the chromatic scale are plotted equidistant order , and it should be remembered that the arithmel increase in the number of revolutions to raise the pitch of the no from C ( 128 vibrations ) to C ( 256 ) , is only half that to raise it fro C ( 256 ) to C ( 512 ) , and so on .
The effect of plotting on the axis Y the actual number of revolutions , would be to make all fc curves a great deal steeper .
It seems necessary , too , to mentif Contributions to the Study " Flicker.'9 t t the writer 's eyes are perfectly normal , and that he has a good There are so many experimental details , that to describe fully angle observation , and to state the reasons for the many precau-v os which must be taken if the result is to be of any value , would e up so much space that the writer thinks it best only to mention ew of the more important:\#151 ; The same spot on the lime was never ; d twice ; the width of the slit was kept unal tered ; each disap-urance of flicker was witnessed by two , sometimes three , observers ; .
, same time as nearly as possible was maintained between the nervations , and used in making them , in order that the retina as 11 as the lime might be in approximately the same condition for \lt ; h ; the room was completely dark* and the eyes rested between the nervations , never looking at any bright objects , and , above all , not the lime ; the oxygen and coal gas were conveyed by special metal pes from gasometers in which they were stored under constant essure outside the building , but since the composition of both ses is liable to variation , the intensity of the light was maintained nstant , by regulating the supply of the gases till the flicker of a If and half disc in the yellow vanished at the same speed of retain ( generally A ' ) .
After every set of experiments , to make sure at the illumination had not sensibly changed , the note for vanish -g flicker in the yellow was again observed .
Experience teaches at the blue-green is rather better than the yellow for this purpose , id accordingly it was used in later experiments .
Not more than w curves , often only one , were drawn in twenty-four hours , and to teck the effects of contrast in consecutive experiments in different .lours , a curve was drawn ( a ) by taking the colours of the spectrum om red to violet and vice versa in their natural order , and by ading the pairs of colours on either side of The yellow for which cker vanished at the same rate o* rotation .
The results of eperiments made thus agree very closely , and the writer may tie here that throughout the eight years during which the research is been carried on , the feeling has steadily grown that in sub-.etive vision we have to do with immutable quantities , capable of met measurement , and which vary but slightly , if at all , with ifferent people .
A glance at the curves in fig. 1 will show that there is a very reat difference in the rates of rotation necessary for the disappear-nce of flicker in the different parts of the spectrum , the speed tor he yellow being very nearly double that for the violet .
Also , that on each side of the yellow the rate decreases , until the 'icker disappears for the last distinctly visible red at flu same iat\lt ; t which it vanishes for the fall green . . .
Pig .
2 presents this last fact in a somewhat different way , and , it * So far as external light is concerned .
Mr. T. C. Porter .
Fig. 2 .
Fe/ s.persec .
for24 holes disc 32* 'Yellow .
-G(re6 ) so -zosFullOrdiyp * -284 -F# \#166 ; .Greenish-Yellow .
\#151 ; f Yellowish-Green .
-\#163 ; -Z6*Vermi/ ion 26 ' 24-240cy^JrFullGreen .
-D 22*21-9 20 .
-ms / a it * 16 " Blue-Green .
~ c# . .
Blue .
f -\#187 ; # -Violet .
-G(S64 ) anything , more clearly ; the yellow and the violet\#151 ; the two extra* so far as the present considerations go\#151 ; are placed at the top .d bottom of the central vertical line , and the number of rotations ?
r second necessary for flicker just to vanish , given in the column* the left , is seen to be ( / or this half-white , half-black disc ) 32 or the yellow , and 16 for the violet , and 24 for the last paiuf colours\#151 ; crimson and full green .
The comparative rates of retain for these colours thus bear the ratios of 2 : 3 : 4 , which are ea , y remembered .
To return to fig. 1 , The curves II and IX lie consideray above the others : they are the expression of observations mas , as suggested to Professor Rood by Dr. Woolcott Gibbs , f viewing the spectrum through a rotating disc having a seqi ' of 180 ' removed , using a telescope .
Though the form of tfc two curves is practically the same as that of the others , tbr higher position , due to the superior intensity of the light receW by the eye in this direct vision method , proves clearly what is alrear well known for white light , that the speed necessary for the disa pearance of flicker increases with increase of the intensity of t light , whatever its colour may be .
No. II is for lime-light , No. I for sunlight , the rays of the sun being reflected through the slit 1 means of a heliostat .
Of the other curves in fig. 2 not much need be said ; they are a for lime-light , the spectrum being thrown on the rotating cardboar disc half white and half black .
The writer has thought it better n( to smooth the curves in any way .
These curves , then , not only confirm the already known fact th^ Contributions to the Study of " Flicker .
" + ; stimulus given to the retina lasts undiminished a shorter time f( the yellow than for any other colour , hut they give the very aaroximately exact relative " last " for the different colours .
The o- eral form of the curves following very closely the curve which presses the luminous intensity of the spectrum obtained by JNivton , Abney , and others . ?
he next point seems to the writer very important , especially with rerence to what follows .
Numerous experiments were made to find o ; how the " last " ( undiminished ) for any one colour varied with t intensity of the light ; for the present it is sufficient to say that isvery case the more intense the illumination the more rapid must r , rotation of the disc be made before flicker will vanish .
Hence v are bound to infer that as the stimulus applied to the retina i reuses in intensity , the impression produced retains its maximum \#166 ; lie for a shorter and shorter time .
Chat a brighter illumination does produce a greater stimulus ( i.e. , 11 neither the contraction of the pupil , nor any other cause , over-cues the effect of brighter illumination ) is conclusively proved by t ) fact that the brighter the light the brighter on the whole is the t.c when flicker has just vanished .
Research was next made to find out in what way the rotation of atisc must be varied for flicker to vanish when the proportion of tj coloured to the black sector varied , the intensity of the illuminate remaining constant .
Two discs were taken : one painted with Indian ink laid on as dry i possible , and the other of the whitest cardboard procurable ; these re dovetailed after Maxwell 's method and mounted on a motor \lt ; th a " siren " disc .
This last was pierced^with six , twelve , or 'enty-four holes , that the note emitted might Re within the octaves uich are both easy to sing and in which intervals can be most \#166 ; erectly judged .
A.s the experiments took many days and nights , and it was most portant that the conditions should be as similar as possible on 'fferent occasions , the whole of the apparatus remained unmoved roughout the whole set of comparative experiments ; the lime-light truer was firmly screwed down to the base-board of the lantern .
ie rotating disc was most carefully screened from extraneous light , ough if the writer repeated the experiments he would wear a black ask to prevent the illumination of the dark sector of the disc by en the faint light reflected from the observer s face .
The wiitei inks this may account partly , though not wholly , for the asym-etry of the curves of fig. 3 in the case of the red and yellow .
The ating was surrounded by a screen coated with the dullest optical ack , whilst another screen allowed only the special colour required fall on the half of the disc viewed .
Each observer regarded the Mr. T. C. Porter .
uogoag 9VMM J ' disc from practically the same stand-point and distance , the W being that of most distinct vision ; moreover , be kept his eyes n\#174 ; on the same part of the disc , so that the image of the disc did \#174 ; travel over the retina .
AJ1 these are important details as will appe later .
Before examining the curves in fig. 3 , consider a disc rotating nod a fixed degree of illumination , i.e. , in a light of constant intensh Suppose that the light is of a certain colour , e.g. , yellow , and t*P * Contributions to the Study of " 353 tain sector only , reflects the yellow light to the eye , the rest of \#187 ; r disc being of a perfect black .
Let the disc be rotating very jdy : then , as the yellow sector passes before the eye , it will , ,.ear as bright as possible under the given illumination , i.e. , it will nsear of no less bright a yellow than it would , if the whole disc n e yellow , i.e. , the stimulus given to the part of the retina on iv ch its image rests , is the maximum possible to this yellow , under 11 particular and constant illumination .
Now , suppose the rate of rxition to be raised until flicker has just vanished , no part of the disc n\lt ; r looks so bright as the sector did at first , and since the final h fitness of the disc varies directly as the width of the yellow s*5or ( this is the fundamental assumption on which all the colour nations rest , and is verified by experiments described later ) , then , wn the flicker has vanished , the effective stimulus at any point of the r , na is to the original and maximum stimulus as the angle of the yellow a , or is to the angle of the whole disc , i.e. , 360 ' , the illumination being distant throughout .
Moreover , since the mere increase in the rate of rotation has no elot whatever on the real width and brightness of the yellow .4tor , but only diminishes the time during which its stimulus is adied to the retina , and diminishes , in the same ratio , the time the bck sector takes to pass ( and the more rapid passage of the black stor must , considered alone , tend to increase the brightness of the \gt ; -c ) , whilst it ( the increase in rate of rotation ) increases to the same \#166 ; ent the number of stimuli applied to the retina per second , and t ' number of transits of the black sector ; it follows , since the final { parent brightness of the disc is less than if it were all yellow , ht the yellow sector requires a finite time in order to produce its maxi-nm effect , and the same argument applies to any colour .
This delusion is in complete accordance with the results of other tperimentalists .
J But this is not all that these considerations prove : for since the iirease in speed of rotation diminishes in the same ratio both the in the image of the yellow sector takes to pass over a point on the i-ina , and also the time the image of the black sector takes to pass e. , the time the sensation evoked by the yellow sector must neces-frily last undiminished , if there is to be no flicker ) , the increased teed would have no effect whatever on th except to multiply e number of times it occurred per second , if it were not that a saker stimulus has a longer " last " ( using the word " last to ean the duration of the sensation undiminished , after the stimulus is been withdrawn ) .
This is a second proof of the principle tablished in a different way earlier in the present paper ( p. 351 ) .
Experiments were next made to measure directly the appaient " ightness of rotating flickerless discs , and to find an expression or Mr. T. C. Porter .
the effect of successive equal increments to the bright sector 'j , method used was to measure the distances from a movable sourc 0f light , of the rotating disc , and of a fixed disc , of the same coloi ag the bright sector of the rotating disc , when the brightness of h , two discs appeared equal .
It was found that for illuminations aJ , ut the same as those used in the other experiments recorded in iis paper , the law which connects the apparent brightness with he width of the bright sector is that enunciated before , .
, th a flickerless half-and-half disc appears half as bright as the fixed ad wholly white or coloured disc , at any rate within the erromf experiment which , however , in this part of the research were of inconsiderable .
When the width of the white or coloured seor was increased in steps of 10 ' at a time , the increment of ae apparent brightness in the flickerless disc followed , within the errs of experiment , the series 1/ 0 , 1/ 1 , 1/ 2 , 1/ 3 , 1/ 4 , \amp ; c. , as it should ; id since these fractions express the ratio between the increased stimulus caused at any stage by an additional 10 ' of colon d sector , and the stimulus already existing before its addition , it folks that we can , with the help of the principles established alrea , ' , predict how the rate of rotation for the disappearance of flicker n t vary with the growth of the white or coloured sector .
We shod expect the first few additions of the 10 ' to produce the most marld alteration ( a rise , as we already know ) in the rate of rotation nec.-sary for the disappearance of flicker , since the diminution in w amount of the black sector is trifling in comparison with its tot width .
Towards the final additions of 10 ' , we shall reach a sfen when the effect of the increment of 10 ' of white or colour is almt negligible in comparison with the total width of the white r coloured sector , but just at this time , the relative diminution of w black sector will be most rapidly , and in order tit flicker may only just be invisible , the rate of rotation mi be considerably diminished , and this diminution will bear to is total velocity almost the ratio the diminution of the black sect ' ( 10 ' ) bears to its total width ; but not exactly this ratio , for sii5 the effective stimulus is still increasing , though the increase is sms compared with its total magnitude , and since this implies\#151 ; aps from any effect of change in the width of the black sector\#151 ; that t * rate of rotation must be raised for the flicker to vanish , it folio ' that the rate of rotation will not diminish at so rapid a pace as ti shrinkage of the black sector 's width would demand , if it alone b* to be considered .
So far , therefore , the " flicker " curve ( for a di with a growing white or coloured sector , the angle of the brig sector being measured on the axis of Y , and the speed of rotation t the axis of X ) will be , on the whole , symmetrical with respect to tl straight line passing through the point on the axis of Y correspon Contributions to the Study of " Flicker .
" " to the half-and-half disc , bnt not completely so , for the con-\gt ; .-i ration just mentioned would cause the curves which are repre-ae ; ed in fig. 3 to be steeper on the right than on the left , viewing f t curves from the axis of Y. It will be seen at once that in prac-ni the converse is true , and the writer believes that this is due to tit fact that the black sector is not completely black .
The effect of i i small percentage of light reflected from the black sector will be itirst to diminish the rate of growth of the speed necessary for li icer to vanish , for it diminishes the contrast between the coloured : ; eor and the black\#151 ; a contrast on which the flicker primarily iltends .
The curves will therefore rise more gradually on the .at , holding the figure as already described .
When there is little 4the black left , there will be proportionately little of the light -octed from it , \#151 ; and if this light be bright enough to have any appre-eible effect , the effect must be to make the decrease of the speed of ration ( necessary to cause the flicker to all but reappear ) more rapid , oause it lessens the effect of the narrow black sector left , giving id impression that all the coloured sector 's light has survived the p sage of the black sector , for a rate of rotation at which , in reality , t art failed to survive , and would have produced flicker if unaided oihe light reflected from the black sector .
, 'hus any want of blackness in the black sector will have con-Mrably more effect during the early part of the growth of the ooured sector , whilst the black sector is very large , than afterwards , at this completely explains the observed departure from symmetry nthe curves constructed from actual observation .
It should be ied that from the way in which the disc is illuminated by the a ctnim , any light reflected from the black ^sector is of the same vDur as that of the bright sector .
If pigment had been used to your the bright sector , and the disc viewed in white light , white l'ht would have been reflected from both the coloured and the i ck sector , and the effect of this would be very much harder to '-plain .
five curves will be seen in fig. 3they are for different colours of b same lime-light spectrum ; each of the capital letters gives the flfcct result of an observation , carefully verified in every case .
The \#163 ; fcual number of rotations per second can be easily found for any lint on the curves by dividing the number of vibrations of the cor-Mponding note found on the axis of X by 24 , the number of holes in b siren disc .
The musical intervals , so far as the diatonic scale is 'acerned , are the true chromatic , and not the equal temperament !
stem .
The information conveyed by the position of any point on one of e curves may be stated as follows , taking , for example , Gr on the '0 ' line of the yellowish-green :\#151 ; .356 Dr. J. W. Capstick .
The excitation of the retina caused by the stimulus of the ish-green light of the lime-light spectrum , reflected from \#171 ; ^v cardboard in 1/ 72 sec. ( i.e. , 160/ 360 of 1/ 32 ) lasts undiminish f 5/ 288 sec. ( i.e. , 200/ 360 of 1/ 32 ) , i.e. , about 1/ 58 sec. Taking the next point G ( 180 ' ) above this last point , we finchai the " last " of the stimulus of the same yellowish-green , apply \lt ; f0\#163 ; 1/ 64 sec. is 1/ 64 sec. Thus , tabulating the results for 'a few points including the above two , Stimulus for Y.G. H 5 ?
applied 1/ 72 sec. lasts 5/ 288 ; \amp ; 288/ 5 x 72/ 1 \#187 ; 1/ 64 " 1/ 64 " 64x64 " 1/ 58 " \#187 ; 1/ 49 " \#187 ; 1/ 40 " \#187 ; 1/ 32 " 1/ 72 " 58x72 1/ 92 " 49 x 92 1/ 120 " 40 x120 1/ 160 " 32x160 =147 =096 =176 =508 =80S = 120 Hence the duration of the impression on the retina undiminM appears to decrease as the time of stimulation increases , thigh within narrow limits of variation one of these quantities is nd ?
inversely proportional to the other .
With regard to the total duration of a luminous impression .
; be writer would point out that nothing has been said in this paper hr few experiments he has made to measure this , lead to the belieftafc it is almost of a different order of magnitude from the time dntng which an impression remains undiminished , and is to be meas**fc by whole minutes rather than by small fractions of a second .
* " On the Kathode Fall of Potential in Gases .
" By J. V. Capstick , M.A. , D.Sc .
, Fellow of Trinity College , Oxbridge .
Communicated by Professor J. J. Thomson , , F#H Received May 17 , \#151 ; Read May 26 , 1898 .
It has been shown by Hittorf* that when an electric current pa ; s through a tube containing a gas at a pressure of a few millimel-t there is a rapid fall of potential near each of the electrodes , wit ; ; much more gentle fall in the space between , and whilst the fall rr the anode and in the positive column varies with the density of c gas and the current strength , the fall near the kathode is const ; ' Warburgf has made careful experiments on the kathode fall , and s fully established its constancy .
If the gas is pure and dry , electrodes clean , and of a metal not acted on chemically by the g\lt ; * * Wied .
Ann. , ' vol. 20 , p. 705 .
t ' Wied .
Ann. , ' vol. 31 , p. 545 j vol. 40 , p. 1 .

