Prof) Forbes 07i the Refraction and Polarization of Heat. 139 



inches, whence we may compute the size of the lunar image 

 to be a circle 0*38 inch in diameter. Comparing this with 

 the dimensions of the intercepted cylinder of rays, we shall 

 find the concentration to exceed 6000 times. But even if we 

 admit that half the rays are reflected, dispersed and absorbed, 

 we shall have still an effective increase of 5iOOO times. 



12. My experiments were made on the 16th December 

 1834, between 9 and 11 o'clock, the moon being only 18 hours 

 past full, and (towards the close) less than 2 hours from the 

 meridian. She was also particularly high, having a declina- 

 tion of 25° north. The thermal pile, which was particularly 

 commodious for the experiment, had one extremity elevated 

 to the proper angle, and being placed accurately in the focus 

 of the mirror, the moon's image was brilliantly thrown on the 

 extremity of the pile. The sky was on the whole very pure, 

 though an occasional milkiness was perceived, but the best 

 observations were made at the clearest moments, because then 

 the air was also most still ; for though the instrument was 

 placed in a most sheltered spot, the faintest breeze was indi- 

 cated by a deflection of the needle, and with such promptitude, 

 that I generally could perceive in this way its approach before 

 I could feel it. The action of the lens was so perfect, that the 

 image was perfectly sharp, and the spots clearly defined. The 

 lunar rays were alternately screened and admitted by an as- 

 sistant passing a sheet of pasteboard across the surface of the 

 lens next the moon -, for when it was interposed between the 

 lens and the instrument, a sensible disturbance took place. By 

 these and other precautions, the needle was steady beyond my 

 expectations, and during an hour and a quarter that the ob- 

 servation lasted, I had probably at least twenty perfectly unex- 

 ceptionable comparative observations, free from the influence 

 of wind, and which invariably gave not the faintest indication 

 of warmth. When I got a deviation of the needle at the mo- 

 ment of unscreening the moon's rays, I verified it by screen- 

 ing them instantly, and watching for a return to zero, but I 

 was always disappointed. I feel quite confident that the effect, 

 if there was any, could not amount to a quarter of a degree of 

 the galvanometer ; and, owing to the dynamical effect which 

 I have described of a first impulse, that it is improbable that 

 it amounted to half that quantity. 



13. Hence it becomes an object of interest to form some 

 estimate of the sensibility of the thermo-multiplier, compared 

 to common thermometers. It would be difficult to give a pre- 

 cise measure of the degrees of temperature of the two extre- 

 mities of the pile*, but we may compare the effect of equal 



* This might best be done by adapting a differential thermometer of ex- 

 treme delicacy, so that the balls might be in contact with the two extre- 



T2 



