308 KADIANT IIEAT. 



might be accounted for, without the necessity of supposing any pe- 

 culiar property of heat, or a direct transmission even through the 

 second glass. 



My apparatus, in following up this inquiry, was similar to that 

 described by M. De la Roche, and consisted of two tin reflectors; in 

 one focus the bulb of a thermometer coated with Indian ink, and in 

 the other an iron ball two inches diameter, which was heated to red- 

 ness, and then cooled till it ceased to be visibly red in the dark, at 

 which point it was placed on its stand, and a thick screen withdrawn. 

 The indications were observed, first, for the direct effect ; secondly, 

 with one glass screen interposed; and thirdly, with two. The tem- 

 perature of the screens was observed by means of a small thermometer 

 attached to the face of each away from the ball, towards its central 

 part; the bulb being kept in contact by the spring of a wire with 

 which the thermometer was fastened. 



The results are: First. That the additional diminution occasioned by 

 the second screen is proportionally much smaller than that occasioned 

 by the first. Thus De la Roche's conclusion is shown to hold good, 

 not only in the case of luminous, but also of non-luminous hot bodies, 

 which is perhaps of consequence, as I believe doubt has been enter- 

 trined respecting it; and it may be remarked that here the greater 

 thickness of the second screen would be against such a result. Secondly. 

 If the progress of the indications of the direct effect bo followed, it 

 appears that the rise in the fii^st 30 seconds is the greatest, and that 

 those in the subsequent periods gradually diminish. Thirdly. With 

 one screen the effect in the first period is equal to or even less than 

 those in the subsequent ones; and. if we follow the temperature of the 

 first screen, it appears to sustain a rapid increase at first, and after- 

 wards continues gradually to rise till some time after the focal ther- 

 mometer has become stationary. The progress of the focal thermom- 

 eter exactly accords with what must be the heating effect of the 

 screen as a source, viz: rising slowly at first as the screen acquires 

 heat sufficient to supply it, and remaining stationary so long as the 

 still increasing temperature of the screen could balance its loss of 

 heat. Fourthly. With two screens there is no rise till the second half- 

 minute, when it is not greater than in the next half, after which the 

 thermometer becomes stationary, and this trifling effect exactly ac- 

 cords with what the temperature of the second screen should produce. 

 It does not begin till the second screen has acquired a higher 

 temperature, and it is stationary while the temperature of the screen 

 continues to increase, and the temperature of the second screen is 

 such as is clearly accounted for from the heating effect of the first. It 

 does not begin to rise till after that of the first has risen; it continues 

 stationary some time after the first has begun to cool, as the first 

 screen did when the iron was cooling. But as in this case the source 

 of heat was cooling during the whole time of the experiment, whilst 

 in the other it was heating during the first part of the time, it follows 

 that a greater proportional temperature should be communicated to 

 the second screen by the first, than to the first by the iron ball. 



Other circumstances will partially co-operate in producing this 



