Nov. M, 1886] 



NA TURE 



45 



burner behind a lens more than a yard square, by placing over 

 it three other similar burne-s and lenses. But it seemed a 

 hazardous experiment to imitate this plan by placing three lamps 

 fed with mineral oil one over the other. However, the skilful 

 arrangements of Sir James Douglass were completely successful. 

 The three superposed oil lamps burnt as safely and well as if 

 each had had the lantern to itself. 



II. — Arran«cmails for Obsirvalion 

 A short account of these arrangements have been given in my 

 previous report, and a complete account is to be found in the 

 report of the Trinity House Committee. 



The plans for making observations on shore at small distances 

 had been well laid. The home at St. Margaret's, stationed 

 between the two observing huts, with telephone to all pjints ; 

 the measured distances ; the huts themselves, welcome refuges on 

 a cold night, and most convenient for photometry with their 

 helpful occupants ; all bore witness to the wise forethought which 

 had been bestowed upon the details of the inquiry. 



For obtaining records of the relative brightness of the different 

 lights from the impressions of those who saw them, probably no 

 better plan could have been devised than that of distributing 

 forms to be filled in with a numerical estimate of the ratio which 

 two of the lights bore to the third. And the enlisting a multitude 

 of observers, by the wide distribution of these forms, secured 

 the two advantages, of an a\erage drawn from a very large 

 number of observations, and of an obviously impartial judgment. 

 The observations made at sea from the Trinity House yacht 

 Ar^us, which was in constant attendance, were of great im- 

 portance ; and 1 may add that, for the landsmen whose main 

 business was photometry at small distances on shore, taking part 

 in these observations was an essential help towards the full 

 appreciation of the problem before them. 



lll.—P/iolomelry 



The chief assistance which I found myself able to render to 

 the Committee was in devising and improving photometric 

 apparatus and methods. A full description of these is given in 

 the Committee's Report, especially in Mr. Di.xon's " Record," 

 part ii., pp. 30-36. During my visits to the South Foreland, 

 I was principally occupied with photometry, in the dark gallery 

 by day, and in one of the huts by night. Frequently Mr. Long- 

 ford or Mr. Dixon worked with me, and the observations which 

 I made are included in the general record. I believe that the 

 standard of light employed was constant and of a definite and 

 reproducible value, and that the methods of comparison were 

 trustworthy and accurate. The excellent idea of Sir James 

 Douglass, of using a large lens to concentrate the rays from the 

 lighthouses upon the photometric disk, m,ade possible the 

 measurement in the more distant hut of lights whose intensity 

 was too feeble to be accurately estimated without such aid. 

 Mr. Dixon's polariscope photometer and the ingenious obscura- 

 tion photometer of Captain Nisbet, are instruments well adapted 

 for the direct comparison of distant lights or lights enfeebled by 

 haze ; the former can only be used for lights which are near 

 together. Two movable photometer-bars, designed by Sir 

 James Douglass, and suitable for use with any form of disk and 

 any standard, were in constant employment throughout the 

 trial. These were placed in the photometric gallery and in hut 

 II. The observations in hut I. were reduced by means &f a 

 portable bar devised and made by Mr. Dixon. 



A glance at the 16 columns of the closely-printed photometric 

 record, — each number bemg, as a rule, the average of many 

 observations, — will give to those who know the effort of atten- 

 tion which accurate photometry requires a conception of the 

 diligence with which this branch of the inquiry was pursued. I 

 have spent many hours in one or other of the huts with Mr. 

 Di.xon or Mr. Longford, and I wish to express my conviction that 

 the results which they obtained and which are printed in the 

 Trinity House Report, are as complete and trustworthy as 

 zealous, patient, and skilful work could make them. 



IV. — Comparison of Lights 

 -Although an observer's opinion on the relative brightness of 

 two or more lights, like an opinion on the force of the wind, is 

 better expressed by means of numbers than by descriptive terms, 

 such numbers must not be regarded as expressing the relative 

 intensity of the lights so compared. Perhaps on any future 

 occasion it would be better to call the brightest light 10 rather 

 than 100, since the use of the larger number suggests that an 



inferiority Just sufficient to be noted with confidence is to be 

 expressed by a difference of 2 or 3 per cent., whereas it probably 

 amounts to at least 10 per cent., and 9/10 or 8/10 would be 

 nearer the ratio of the two lights than 98/100 or 95/100. No 

 doubt by practice in comparing lights whose relative intensity is 

 known, a fair power of judging may be acquired ; but without 

 such training the natural tendency is to under-estimate differ- 

 ences. For example, the average of 294 estimates by eye of the 

 relative power of " triform oil " and " quadriform gas," assigns 

 to the gas a superiority of 6 per cent., whereas the actual 

 superiority as shown by measurement is 23 per cent. According 

 to the same series of estimates the electric light has a superiority 

 over "quadriform gas" of 59 per cent., the actual superiority 

 being more than 400 per cent. 



Equally remarkable evidence of the tendency to overlook 

 differences of intensity when the estimate is made directly by eye, 

 is found on comparing the values assigned to "multiform" 

 lights. The figures relating to the loSjet gas-burners, and 

 representing the relative value of the single, biform, triform, 

 and quadriform lights, each by comparison with the electric 

 light, are 56, 61, 59, and 6j. If these numbers represented the 

 intensity of the light falling upon the eye from the whole surface 

 of the illuminated lenses, they should stand in the ratio of I, 2, 

 3, and 4. The explanation at once suggests itself that while the 

 photometer measures the total light received from a large 

 illuminated surface, the eye judges of the brightness of the 

 surface or the light received from equal areas. To an observer 

 looking down a street on a clear night, the more distant gas- 

 lights seem as bright as those which are nearer, though smaller 

 in size ; if asked to estimate the lights he would probably assign 

 the same figure to all. And the visual angle subtended by the 

 flame of a street-lamp at too yards is about the same as that sub- 

 tended by 18 feet of lenses at a distance of 4 miles. The 

 singular fact that as estimated by eye, on a separate comparison 

 with the electric light, " multiform " have no superiority to single 

 lights, may to a small extent admit of the explanation which 

 applies to the familiar case which has been given. In this case 

 the observer distinguishes between size and brightness, and sets 

 himself to judge of the latter only ; or it may be that the intensity 

 of the sensation of light depends upon the brightness and not the 

 size of the spot of light formed upon the retina. But the South 

 Foreland observations were not made chiefly at distances of only 

 2 or 3 miles ; nor were the observers likely to disregard the 

 apparent magnitude of a light in estimating its value. At dis- 

 tances of from 12 to 14 miles, at which the largest lights have no 

 appreciable magnitude, the average values assigned to the biform, 

 triform, and quadriform lights are 75, 66, and 60. 



I fear the true explanation is that the results have suffered 

 from the electric light having been adopted as the term of com- 

 parison. To a small extent its fluctuations and difference of 

 colour, and to a much greater extent its incomparable power, 

 have made the estimates entirely uitcertain ; and thus it is vain 

 to institute cross comparisons between the different lights which 

 were not seen together, but only estimated by reference to the 

 electric light as a standard. This conclusion, however, does not 

 affect the value of the comparisons, chiefly aimed at, between the 

 gas and oil lights, which were seen together, nor the significance 

 of the direct comparison of the flame lights with the electric 

 light at all distances and in all weathers. Indeed, the adoption 

 of the electric light as the standard with which all others were to 

 be compared, has served to establish on the basis of thousands of 

 observations the important fact that, as far as the eye can judge, 

 the electric light appears to excel the light of gas or oil lamps 

 almost as much at greater as at smaller distances, and in hazy 

 weather as in clear. The mean ratio of the electric li-;ht to all 

 the gas and oil lights exhibited, taken from the whole number of 

 recorded observations, is, at distances of from i to 8 miles 

 1000/626, at distances of from 8 to 15 miles 1000/613. In clear 

 weather the mean ratio is 1000/591 ; in weather not clear it is 

 1000/608. 



The photometric record presents many points of interest. 



In the measurement of naked flames the long gallery, which 

 has wisely been made a permanent structure at the South Fore- 

 land, afforded unexampled facilities. The electric arc was 

 measured at a distance sufficiently great for its intensity to be 

 similar to that of the other lights which were measured. The 

 values assigned to it, from 10,000 to 15,000 candles, are not so 

 high as some which have been obtained ; but this is perhaps due 

 to the fact that extreme values were rejected, and care was taken 

 to obtain an average result. 



