16 



PROFESSOR FORBES'S RESEARCHES ON HEAT. 



33. The numbers in the last column (computed on the supposition of the 

 interstices being geometrical rectangles) are to be compared with the following 

 experimental determinations of the proportion of incident heat transmitted by 

 these gratings. 



Proportions per 100 of Incident Rays of Heat transmitted by Wire-Gauze. 



The differences for each grating, perhaps, do not exceed the errors of experiment. 

 In every case these numbers are inferior to the geometrical interstices, but what 

 inclines me to think that this difference is due to the irregularities of figure of the 

 gauze (including (he effect of flattening of the wires where they overlap, making 

 the interstices obtuse-angled) is this : that No. 2, in which the wires were finer 

 compared to the interstices than in others (the total interstices being one-third 

 part larger in proportion), and the gauze evidently far more regularly formed 

 than in the other cases, the per-centage transmitted differs very little from the 

 geometrical gauge. I own, at the same time, that a difference of 5 per cent, in 

 No. 3 (which is evidently not due to an error of observation), seems to me barely 

 accounted for by this remark. 



34. Thread Gratings. With gratings of fine cotton-threads ^ inch apart, 

 used for shewing FKAUNHOFER'S Spectra, I obtained a similar result. These threads 

 were arranged parallel-wise on two frames, capable of being superimposed rec- 

 tangularly. Thus, we can either employ a screen of parallel threads one-hundredth 

 of an inch apart, or a screen of mathematically accurate squares, formed b} T 

 superposition. It is difficult in this case, however, to obtain the diameter of the 

 thread accurately enough to estimate the ratio of interstices. 



Per-centage of Incident Heat transmitted by Cotton-Thread Gratings, 



^ inch apart. 



