July 21, 1892] 



NATURE 



285 



From this point it would be a very easy step to test the 

 " actinometric effect on sensitized paper" ("chemical action " 

 of Roscoe) of different skies, or of the sun at different altitudes. 

 It is not probable that the chemical action is simply proportional 

 to the light ; but it would be soon found that the " chemical 

 action" could be much more accurately measured than the 



Sir Henry Roscoe (partly in junction with Bunsen and with 

 Thorpe) made many investigations and various publications 

 between 1859-70 on the chemical action of the sun and sky 



< measured by its effect in darkening photographically sensi- 



ed paper. Roscoe delivered the Bakerian Lecture in 1865, 

 ■ On a Method of Meteorological Registration of the Chemical 

 Action of Total Daylight." 



Throughout his investigations Roscoe pursued a direct method 

 of experiment : he elaborately investigated a method for ob- 

 taining always paper of standard sensibility ; he devised a plan 

 for obtaining a light of standard intensity ; he then exposed a 

 piece of the paper to the action of the sky, or of sun and sky, 

 ur of a portion of the sky, and compared the effect produced in 

 a given number of seconds with that produced in the same 

 paper in the same number of seconds by his standard light. 

 Roscoe also, by a laborious method, verified his fundamental 

 asumption that light of intensity 50 acting for i second has the 

 same effect as light of intensity i acting for 50 seconds. 



Roscoe took half-hourly readings at Manchester, and thence 

 gave the (comparative) actinic effects of the sky at different 

 seasons of the year. Also he compared the chemical intensity 

 of total daylight at Kew and Para, and investigated the relation 

 between the sun's altitude and the chemical intensity of total 

 daylight in a cloudless sky. By total daylight Roscoe meant 

 the chemical action produced by the sun and whole sky together 

 on a piece of paper exposed horizontally. 



Roscoe found that his readings were enormously affected by 

 the cloud-haze or invisible vapour in the air in England ; he 

 got his results, as to comparison of the chemical intensity at 

 different seasons of the year, and at different altitudes of the 

 sun, by assuming that in the average of a large number of 

 observations, the effects of cloud, &c., would be self-destructive. 



Roscoe found that the " chemical effect " of the sun depended 

 only on his altitude (in a cloudless sky), being the same at Para 

 and at Kew. He got very anomalous results as to the effects in 

 spring and autumn in England, probably because the effects of 

 cloudiness were not self-destructive in his series of observations. 

 He arrived, by " averaging" the cloud irregularities, at the law 

 that "the relation between the sun's altitude and the chemical 

 intensity of total daylight is graphically represented by a right 

 line " (a result only a rough first approximation to the truth). 

 Roscoe obtained small result in comparing the chemical action 

 at different points of the same sky, partly because he could make 

 no experiments in person on a tropical clear sky, partly because 

 to note these differences requires superior instruments to the 

 direct experiment method alone tried by Roscoe. 



Mr. W. Brennand was engaged at Dacca in observations, 

 jiarallel to those of Roscoe, and nearly contemporaneous, 1861- 

 66. Brennand was quite unaware of Roscoe's experiments. 

 I'eing an amateur photographer, and his own photographic 

 chemist, he was first led to devise an instrument for testing the 

 chemical action of sun and sky, in order to obtain guidance for 

 the number of seconds to expose a photographic plate. He 

 was soon led on to investigate the effect of the sun at different 

 altitudes, the effect of the sky for different altitudes of the sun, 

 and finally the law of distribution of the " chemical action" in 

 a perfectly cloudless sky. 



Brennand's procedure in experiment differed fundamentally 

 rom Roscoe's in two points : — 



(i) Brennand only attempted observation in the cold weather 

 at Dacca when he had a complete horizon of clear sky. He was 

 thus enabled to carry his investigations into the laws of chemical 

 action in a cloudless sky much farther than Roscoe, 90 per 

 cent, (at least) of whose observations were obscured by cloud 

 irregularities that could not be allowed for. 



(2) Instead of Roscoe's direct method of observation, 

 lirennand was early led to devise an instrument (the water- 

 motion actinometer (see Nature, Januarys, 1891, p. 237), by the 

 aid of which he was independent both of the standard light and 



mdard paper attained by Roscoe with so great labour. The 



II himself was, in fact, Brennand's standard light, and the 



irkening of each paper was read as a ratio ; for instance, if an 



exposure of 10 seconds to sun and sky produced the same 



tint in the paper that was produced by the sun alone in 17 



NO. II 86, VOL. 46] 



seconds, then the effect of the sun alone was reckoned \^ of 

 the sun and sky together. It is clear that any uniform paper 

 should give such ratios the same, though the actual shades pro- 

 duced would be different in different papers. All the papers 

 made by Brennand himself were found "uniform," i.e. to 

 within the limits of variation (say, 2 per cent.) within which the 

 darkened paper can be read, i.e. the shades can be matched. 

 Any good photographic paper is found uniform enough for the 

 purpose ; but some of the ordinary photographic papers tried 

 lately in England have been found not good enough ; the 

 nature of the irregularities introduced by imperfect paper is such 

 as to suggest very soon their cause. 



It is to be noticed that all that can be observed is a ratio : 

 the observations in Roscoe's direct process are not absolute. In 

 that process there is a standard unit, viz, the blackness pro- 

 duced in the standard paper by the standard light action at the 

 unit of distance for n seconds. Any other light that produces 



this blackness has the numerical value - in Roscoe's unit. 



n 

 There is little doubt but that Roscoe got his standard light 

 and standard paper, each time he recovered them, correctly 

 within the percentage of error involved in the reading. He 

 would be certain to have prepared his salts of exactly the proper 

 strength ; but there is an element of uncertainty in the degree 

 in which papers apparently of similar texture and in a similar 

 state of dryness, &c. , take up salts. This element of uncertainty 

 is avoided by Brennand's method, which is far more absolute 

 than Roscoe's. 



The water-motion actinometer gave Brennand, for each 

 observation, a shaded strip darkened gradually from o to 8 (or 

 to 16) seconds. He could note on this the point at which a 

 particular unit of darkening was produced, and the inverse of 

 this time gave him a measure of the ratio of the observed 

 " chemical action " to that which had produced the unit 

 darkening. 



This, of course, involved the assumption that light of in- 

 tensity 50 acting for i second has the same effect as light of 

 intensity i acting for 50 seconds. This Brennand thought might 

 be assumed ; but he proved it in the following very simple 

 manner. 



A slip of sensitized paper is formed into a ring (a short 

 cylinder) and placed round a light (the wick of a candle was 

 used, but any light would do, irregular or not) excentrically. 

 After a certain time the slip is examined and found to be shaded 

 gradually from the farthest to the nearest point, the effect at 

 each point varying inversely as the square of the distance. 



Thus if A be the source of light, O the centre of the ring, and 

 if we have OB = a, OA = b, POB = fl, we shall have the 

 chemical effect at any point P of the slip vary as 



I 



AP2 



- 2ab cos 0'. 



In a particular experiment Brennand took a = 1*4 inch, 



b =■ '\ inch. 



o fl 3-24 - R2 



cos^ - = ^-—^ . 



2 2*24 



Taking the unit of intensity that at the distance i inch from 

 the wick, and calculating the values of Q for values of the in- 

 tensity I, 75, -5, and -3, we have Q - o, 20° 10', 78° 34', and 

 141° 48' respectively. The lengths of arc corresponding to these 

 are found to be -49 inch, 1*92 inch, and 3-45 inches respect- 

 ively. These lengths can be marked off on the slip. Another 

 slip can then be darkened in the water-motion actinometer, by 

 any light ; a unit can be marked on this slip at the point where 

 the shade corresponds with that at the unit in the ring slip ; it 

 then can be seen whether the intensity of shade at the distance 

 •49 on the ring slip agrees with that at three-fourths the time 

 for unit on the actinometer slip ; and similarly for the other cal- 

 culated values. This experiment verifies the law assumed, and 

 moreover affords a check on the paper employed, and on the 

 closeness with which tints can be matched. 



Another important means of verification was employed by 

 Brennand, which Roscoe does not appear to have availed him- 

 self of. Calling the effect of the sky alone in darkening paper 

 B, and the effect of the sun and sky together A, Roscoe observed 

 A and observed B, and then calculated the effect of the sun alone 

 as A - B. Brennand did this ; but also observed the sun alone 

 by the simple device of a vertical slit in a shutter, and was thus 

 able to check the accuracy of his method and of his work. 



Having thus established the trustworthiness of his modui 



