126 Dr. S. P. Thompson on Photometry. 



are in consequence of the absolute magnitudes of the standards 

 habitually employed. If a six-foot bench is the right sort of 

 apparatus to satisfy the needs of the eye when the standard 

 is of one or two candle-power, obviously a longer bench will 

 be right when standards of greater power are employed, as 

 giving, on the average, similar absolute illuminations of the 

 working surfaces. If the degree of accuracy implied in the 

 possibility of reading to within 2 millimetres'' length of scale 

 can be attained with a Bunsen or Jolly photometer when the 

 standard is a 2-candle Methven slit, or a pair of standard 

 sperm-candles, an accuracy of five times as great ought to be 

 attainable if there is used as a standard a 50-candle light. 

 The bench need not in this case be five times as long as the 

 ordinary bench : the half length only need be elongated five- 

 fold on one side of the screen, the other half remaining as 

 before. In other words, if the standard is of 25-fold brilliancy, 

 it must be placed 5 times as far away from the screen as 

 before in order to balance a given light at the same distance 

 as before on the other side. And, so far as such standard 

 light is concerned, a five-fold accuracy will be attained, since 

 any error or uncertainty in reading the scale will be now but 

 a fifth part of the whole scale-reading. 



All the foregoing points, then, to the use of a brighter 

 standard light and a longer photometric bench than hereto- 

 fore. Such a standard might well be afforded by an arc 

 crater viewed through a circular aperture 1 millimetre in 

 diameter, giving about 0*7854 sq. millimetre of crater-surface, 

 with a light of about 55 candles. This should be placed at 

 one end of a bench some five metres in length ; the gradua- 

 tions of the scale being, of course, reckoned from the edge of 

 the aperture *. 



One not unimportant advantage of the use of such a pin- 

 hole standard is that it may with real propriety be treated as 

 a luminous point ; whereas no one can maintain that the 

 flame standards habitually used are even approximately points 

 relatively to the distances at which they are set from the 

 screen. For these the law of inverse squares cannot possibly 

 be true ; though it is never the practice to make any correc- 

 tions for the errors arising from the size of the flame. 



In contemplating the use of circular apertures pierced in 

 metal diaphragms, it becomes necessary to inquire how far 



* It is curious to note in this respect that there is usually an erroneous 

 instruction observed with the use of the Methven slit ; the distance of the 

 photometric screen being reckoned from the flame behind the slit instead 

 of from the slit itself. 



