of Heat by Roughness of the Surface. 447 



out, in this scooping a pressure is exerted upon the adjacent 

 parts, which, even though exerted laterally, must necessitate a 

 condensation. 



Melloni* also made the following experiment. From a sheet 

 of plate glass 11 millims. in thickness he had four pieces 

 cut ; they were heated to redness, and two were allowed to cool 

 slowly, while the others were cooled rapidly. One of each of 

 the two kinds was scratched, and then the four sides of a box 

 were made of the plates, which was filled with hot water. The 

 two slowly cooled ones exhibited the same radiation. Of the 

 quickly cooled ones, the scratched one gave a deflection of 29°*7, 

 and the other of 28°. Thence Melloni concludes that the rough- 

 ness of the surface only exerts an influence if the interior of the 

 mass has a smaller density than the radiating layers on the 

 surface. 



Knoblauch f subsequently made experiments to confirm in 

 another manner Melloni's view. He used cast and rolled lead 

 plates. One of the first, the cast ones, which when smooth pro- 

 duced a deflection of 49°, exhibited after being scratched only a 

 deflection of 48°'25. Hence the radiation had decreased — it is 

 true, only by o, 75 ; but by a subsequent scratching in a trans- 

 verse direction it decreased by another degree, for the deflection 

 only amounted to 47° '25. 



Of the rolled plates, one, which when smooth had produced a 

 deflection of 50 0, 5, after scratching also indicated a decrease in 

 the radiation, for the deflection amounted only to 48°*5. After 

 being again scratched in a transverse direction there was again an 

 increase in the radiation of 1°'5 ; for the deflection amounted to 

 49 0, 75. M. Knoblauch supposes that the increase in the radia- 

 tion by the second scratching might be due to the lead being 

 compressed in the positions of the marks, but loosened at the 

 points at which the upturned edges of the furrows intersected. 



Knoblauch then took a cube of rolled copper plate, and first 

 compared the radiation of a smooth side with one scratched in 

 two directions. The first produced a deflection of 29°, the 

 other of 47°* 75. Afterwards, both having been covered galva- 

 noplastically with copper, the first gave 49°*25 5 the grooved one 

 51°5. 



This experiment is indeed very interesting ; but in my opinion 

 it does not, as Knoblauch maintains J, confirm Melloni's prin- 

 ciple " that scratching the surface only influences the radiating- 

 power of bodies so far as it modifies their density and hard- 

 ness, and it increases or diminishes the radiating-power accord- 

 ingly as it loosens or condenses the parts affected;" for M. 



* Thermocrose, p. 88, note. 



f Pogg. Ann. vol. lxx. p. 343. J hoc. cit. p. 343. 



