278 
NATURE 
[ Han. 18, 1883 
shorter than one four-bundred-million-millionth of a second, and 
longer than one eight-hundred-million-millionth of a second, the 
radiation is perceived as light, by the eye. 
Pouillet, from a series of experiments, deduced a value 
of the energy radiated by the sun, equal in British units to 
about 86 foot-pounds per second per square foot at the earth's 
surface, or about 1 horse-power to every 64 square feet of the 
earth’s surface. We may estimate from this the value of the 
solar radiation at the surface of the sun. The sun is merely an 
incandescent molten mass losing heat by radiation, and sur- 
rounded by an atmosphere of incandescent vapour, so that the 
radiant energy really comes out from any square foot or square 
mile of the sun’s surface, as from a pit of luminous fluid which 
we cannot distinguish as either gaseous or liquid. Take, how- 
ever, instead of the sun, an ideal radiating surface of a solid 
globe of 440,000 miles radius. The distance of the earth being 
taken as 93 million miles, the radius of the sun is equal to, say in 
round numbers, one two-hundredth of the earth’s distance, hence 
the area at the earth’s distance corres; onding to one square foot 
of the sun’s surface, is equal to 40,000 square feet. The radia- 
tion on this surface is (40,000 X 86, or) 3,440,000 foot-pounds, 
which is therefore the amount of radiation from each :quare foot 
of the sun’s surface. This amounts to about 7000 hor:e-power, 
which, according to our brain-wasting British measure, we must 
divide by 144, if we wish to know the radiation per square inch 
of the sun’s surface, which we thus find to be 50 horse-power. 
_ The normal current through a Swan lamp giving a 20-candle 
light is equal to 1.4 amperes with a potential of 40 to 45 volts. 
Hence the activity of the electric working in the filament is 61°6 
ampere-volts or Watts (according to Dr. Siemens’ happy desig- 
nation of the name of Watt, to represent the unit of activity 
constituted by the amrere-volt). To reduce this to horse-power 
we must divide by 746, and we thus find about 1-12th of a 
horse-power for the electric activity in a Swan lamp. The 
filament is 34 inches long, and ‘or of an inch in diameter of 
circular section ; the area of the surface is thus 1-9th of a square 
inch, and therefore the activity is at the rate of 3-4ths of a 
horse-power per square inch, Hence the activity of the sun’s 
radiation is about sixty-seven times greater than that of a Swan 
lamp per equal area, when incande:ced to 240 candles per horse- 
power. 
In this country the standard light to which photometric 
measurements are referred is that obtained from what is known 
as a standard candle. Latterly, however, objections have been 
raised against its accuracy. It has been said that differences of 
as much as 14 per cent. have been found in the intensity of the 
light given by different standard candles, and that serious 
differences have been observed in the intensity of the light from 
different parts of the same candle in the course ofits burning. The 
Carcel lamp, the standard in use in France, has been regarded 
as the only reliable standard. It is, no doubt, very reliable and 
accurate in its indications, but it should be remembered that its 
accuracy is greatly owing to the careful method and the laborious 
precautions taken to secure accuracy. If something akin to 
the precautions applied to the Carcel lamp by Regnault and 
Dumas were ap} lied to the producticn and use of the standard 
candle, there is little doubt but that sufficient accuracy for mcst 
practical purpcses could also be obtained with it; probably 
= good results as are already ol tained by the ure of the Carcel 
amp. 
At the Conference on Electrical Units which met in Paris 
lately, a suggestion was made to ue as a standard for photo- 
metric measurements the incandescence of n:elting platinum, 
and very interesting results and methods in connection with the 
proposal were presented to the meeting. According to experi- 
ments by Mr. Violle, which M. Dumas reported to the Con- 
ference, a square centimetre of liquid platinum at the melting 
temperature gives of yellow light seven, and of violet twelve 
times the quontities of the same cclours given by a Carcel lamp. 
The apparent area of the Swan filament, beins one-ninth of a 
square inch, is "23 of a square centimetre, and when incandesced 
to 20 candles must be about as bright as the melted platinum of 
Mr. Violle’s experiment, as the 7 carcels of yellow and 12 of 
violet must correspond to something like 10 carcels or 85 candles, 
in the ordinary estimation of illumination by our eyes. The tint 
of Mr. Violle’s glowing platinum cannot be very different from 
that of the ordinary Swan lamp incande:ced to its ‘* 20 candles.” 
Thus both, as to tint, and brightness, it appears that melted 
platinum at its freezing temperature is nearly the same as a carbon 
filament in vacuum incandesced to 240 candles per horse-power. 
For approximative photometric measurements the mo:t con- 
venient method is certainly that of Rumford, by a comparison 
of the shadows cast by the sources of light on a white surface. 
The apparatus necessary are only a piece of white paper, a 
small cylindrical body such as a pencil, and a means of mea- 
suring distances. Ordinary healthy eyes are usually quite ccn- 
sistent in estimating the strength of shadows, even when the 
shadows examined are of different colours, and with a reasonable 
amount of care photometric measurements by this method may 
be obtained within 2 or 3 per cent. of accuracy. The difference 
in the colours of the shadows is of course due to each shadow 
being illuminated by the other light. 
Arago has compared the luminous intensity of the sun with 
that of a candle, and estimates it as equal to about 15,000 times 
that of a candle flame. 
Seidel, as Sir W. Thomson had been informed by Helmholtz, 
estimated the luminous intensity of the moon as about equal to 
that of grayish basalt or sandstone. An experiment on :unlight 
made in Glasgow on the 8th of this month (since this paper was 
read), compared with an observation on moonlight, which he 
made at York during the meeting of the British Association 
there in 1881, had led him to conclude that the surface of the 
moon radiates something not enormously different from one- 
quarter of the light incident uponit. It would be exactly this if 
the transparency of the Glasgow noon atmosphere of December 
8, 1882, had been exactly equal to that of the York midnight 
atmosphere of September, 1881, referred to below, for the 
respective altitudes of the sun and moon on the two occasions. 
The observation on moonlight referred to above showed the 
mocnlight at the time and place of the observation (at York 
early in September, 1881, about midnight, near the time of full 
moon) to be equal to that of a candle at a distance of 230 centi- 
metres. The moon’s distance (3°8 x 1o!°cm.) is 1°65 x 10% 
times the distance of the candle. Hence, ignoring for a moment 
the loss of moonlight in transmission through the earth’s atmo- 
sphere, we find (1°65 x 10%)”, or 27 thousand million million as 
the number of candles thst must be spread over the moon’s 
earthward hemisphere painted black, to send us as much light 
as we receive from her. Probably about one and a half times 
as many candles, or say forty thousand million million wuld be 
required, because the absorption by the earth’s atmosphere may 
have stopped about one-third of the light from reaching the 
place where the observation was made. The moon’s diameter 
is 3°5 X 10° centimetres, and therefore half the area of her 
surface is 19 X 10! square centimetres, which is nearly five 
times forty thousand million million. Thus it appears that if 
the hemisphere of the moon facing the earth were painted | lack 
and covered with candles standing packcd in square order 
touching one another (being say one candle to every five square 
centimetres of surface), all burning normally, the light received 
at the earth wou!d te abcut the same in quantity as estimated by 
our eyes, as it really is. It would have very n uch the same tintand 
general appearance as an ordinary theatrical moon, except that 
it would be brightest at the rim and cou.tinucusly less bright 
from the rim to the centre of the circle where the brightness 
would be least. 
The luminous intensity of a cl.udy shy he found about 1oa.m., 
one day in York during the meeting of the British Association 
to be such that light from it thr ugh an aperture of one square 
inch area was equal to about one candle, ‘he colour of its shadow 
compared with that from a candle was as deep buff yellow to 
azure blue, the former shadow being illuminated by the candle 
alone, the latter by the light coming through the inch hole in the 
window shutter, ) 
The experiment on sunlight of last Friday (Cecember $) 
showed, at 1 o’clock on that day, the sunlight reaching his 
house in the University to be of such brilliancy that the amount 
of it coming through a pinhole in a piece of paper of ‘09 of a 
centimetre diameter produced <n illumination equal to that of 
126 candles. ‘This is 6-3 times the 20-candle Swan light, of 
which the apparent area of incandescent surface is *23 of a 
square centin etre, or 3°8 times the area of the pin-hole. Hence 
the : un’s surfece as seen through the atmos here at the time ard 
place of observation was 24 times as bright as the Swan carbon 
when incar.desced to 240 candles per horse-power. By cutting a 
piece of paper of such shape and size as just to eclipse the flame 
of the candle and measuring the area of the piece of paper, he 
found about 2°7sq. centims. as the corresponding area of the 
flame. This is 420 times the area of the pin-hcle, and therefore 
he intensity of the light from the sun’s disc was equal to 
