OF ARTS AND SCIENCES. 249 



the heat can certainly not be more than 90 times as great, and is 

 probably much less. Since bodies begin to glow at about 500° C, the 

 following equation will determine the solar temperature: — 



90 (3,500 — 500) -f 500 = 270,500° C. 



Our upper limit would thus be brought down from several millions 

 of degrees to about 270,000° C. 



Now as to the lower limit. The temperature of the hottest blast 

 furnaces is about 2,000° C, or about that of the lime light. That the 

 sun is far hotter than this, or even the electric light, is manifest by 

 an examination of the curves in Fig. 1. Let us take 8,000° as a 

 lower limit, as found by inspection. On observing the spectrum of 

 melted silver, I found that it just about reached to the violet lays. 

 Also the heat of the oxyhydrogen jet is approximately the melting 

 point of platinum. Let us now construct a curve, Fig. 2., in which 

 the unit of abscissas shall be 1,000° C, and the ordinates the same as 

 in Fig. 1., but on a different scale. Then the point Si will repre- 

 sent the position of melted silver, L the lime-light or melted plati- 

 num, and E the electric light. We find that these three points all 

 lie in a straight line. Then if the temperatures we adopted were 

 correct, this would give us a very simple empirical law, viz : — The 

 temperature is always proportional to some function of the ratio of 

 any two assumed wave-lengths. For artificial sources, for the wave- 

 lengths 585 and 455, it varies directly as this ratio. Supposing this 

 law to be uniformly true, the temperature of the sun would be 

 11,000° C. But from a comparison of the experiments of Dr. Vogel, 

 and Prof. Pickering, it would seem that the sun's atmosphere absorbs 

 a much larger proportion of the violet rays, than it does of the yellow. 

 We know this to be the case with our atmosphere, therefore let us 

 double the temperature (and this coefficient cannot be very far out of 

 the way), and we may therefore conclude that the temperature of the 

 sun is approximately 22,000° C. 



This amount is, we notice, considerably within the limits we had 

 previously set. 



Upon this principle, the temperature of the magnesium light, 

 perhaps the highest terrestrial temperature we have yet attained, 

 would be 4,900° C, as shown by Fig. 2. Its small intrinsic bril- 

 liancy is readily accounted for, when we recollect that this depends 

 on the area of the ignited solid matter, and that this, in the case 

 of the magnesium light, consists almost wholly of the impalpable 

 oxide which forms the smoke. 



