250 Messrs. W. E. Wilson and P. L. Gray .
[ Mar. 15 , II .
" Experimental Investigations on the Effective Temperature of the Sun , made at Daramona , Streete , Co. Westmeath .
" By W. E. Wilson , M.R.I.A. , and P. L. Gray , B.Sc. , Assoc. R.C.S. , Demonstrator of Physics , Mason College , Birmingham .
Communicated by G. JOHNSTONE Stoney , F.R.S. Received January 4 , 1894 .
( Abstract .
) The only tolerably complete series of investigations on this subject up to the present time have been those of Rossetti and Le Chatelier .
The results given by other writers have depended more or less on guesses relative to the law connecting radiation and temperature , and differences on this point alone have given values varying between 1500 ' and 3,000,000 ' to 5,000,000 ' C. Rossetti worked with a thermo-pile , exposed directly to the heat of the sun ; the law connecting the deflections of the galvanometer with the temperature of an artificial source of heat having been obtained up to a temperature of about 2000 ' C. , from the deflection produced by the heat of the sun the solar temperature was calculated by extrapolation .
Le Chatelier worked on an entirely different principle , measuring the intensity of the light transmitted through a certain piece of red glass , first from sources at known temperatures up to 1700 ' or 1800 ' , and , secondly , from the sun , the temperature of which was then obtained , as in Rossetti 's case , by a process of extra-polation , which is , of course , necessary in any method , until we can raise substances to a temperature actually as high as that of the sun , an experiment at present impossible .
Rossetti obtained finally a temperature of 10,000 ' C. , approximately , while Le Chatelier gives 7600 ' ( + 1000 ' ) as the mean of his results .
In the paper the difference between Rossetti 's result and our own ( 6200 ' C. ) is discussed , and a possible explanation given .
The method adopted by the authors is a zero method , and the essential point is the balancing of the heat from the sun with that from a platinum strip heated to a high known temperature .
The artificial source of heat was a modified form of Joly 's " meldo-meter , " the calibration of which can be performed with a very high degree of accuracy .
The " radiation balance " is a form of Boys 's radio-micrometer containing a duplex circuit , so designed that the heat from the sun can be made to exert a turning moment in the opposite direction to that due to the artificial source of heat , and by making the apparent area of the latter sufficiently great , the radia1894 .
] On the Effective Temperature of the Sun .
251 tion from it may be increased so far as to equal that arriving at the radio-micrometer from the sun .
The following points are considered , after descriptions of the method and apparatus have been given :\#151 ; 1 .
The law connecting radiation and temperature .
This is probably the most important factor in the value of the final result .
Numerous investigations on the point have been made , which are referred to in the paper .
After a careful series of experiments we have come to .
the conclusion that ( at least for bright platinum ) Stefan 's " law of the 4th power " holds , * i.e. , that for high temperatures ( say over 600 ' or 700 ' C. ) if R = the radiation from a source whose absolute temperature is T , then RocT4 .
2 .
The emissive power of platinum at high temperatures compared with that of lamp-black .
On this point the value obtained by Rossetti was used , some considerations being given in support of his figures .
3 .
The amount of the atmospheric absorption .
This is fully discussed , and again the value obtained by Rossetti is used .
Langley 's theoretical value for percentage absorption of radiation from a body in the zenith , viz. , 41 per cent. , is shown to be possibly too great ; Rossetti obtained 29 per cent. , which appears to be the value best supported by experiment .
The climate of Ireland entirely prevents a systematic series of investigations on this particular point .
Several subsidiary questions are also discussed , and , finally , the results of about sixty-nine observations are given , which lead to a final mean result for the effective solar temperature of 6200 ' C. It is pointed out , in conclusion , that the method would probably give excellent results if adopted in some country in or near the tropics , where atmospheric conditions can be trusted to remain more constant for some weeks , or even days , together , and where a series of observations taken at the same part of the year throughout the period of a sun-spot cycle might be hoped to settle the question of how ( or if ) the solar temperature varies during this time , as any error in the absolute value obtained may probably be considered constant , so that comparative values from year to year might be trusted to indicate any change .
* Stefan , however , stated the law as applying to the " pure " radiation of a surface of perfect emissive power .

