VOL. tx.} a'I philosophical transactions. 73 



emission of light, is what I have made it ; at least it does not exceed what I have 

 made it, be the figure of the particles what it will ; and my conclusions, which 

 rest entirely on the maximum of the solid content, and that of the matter, and 

 are the stronger the less these maxima be, will still hold good. But if these 

 conclusions stand, the objections moved by Dr. Franklin vanish. 



The only part of my reasoning which will be affected, by supposing the figure 

 of the particles of light not to be spherical, will be that in which I attempt to 

 show, in what proportion the magnitude of each particle, and the matter con- 

 tained in each emission, must be less than the maximum, in order to make the 

 density of light no greater than may be consistent with the appearances of na- 

 ture. Here indeed the figure of the corpuscles is of great importance. The di- 

 minution necessary will be very different in different figures, and the figure, I 

 confess, may be such, as to make it much less than what I have shown to be 

 requisite on the spherical hypothesis. However, if ,0006 P^rt of the density of 

 our air be admitted to be as great a density as can reasonably be allowed to light, 

 at the surface of the orbis magnus, the matter of each emission must not, on any 

 hypothesis of the figure of the corpuscles, exceed -j-jVt of the maximum. 



For in order to bring the density down from the maximum, to any other given 

 limit, either the matter must remain unaltered, and the space, which it is sup- 

 posed to occupy, be increased, in the proportion in which the density is to be di- 

 minished; or the space must remain unaltered, and the matter be diminished in 

 proportion to the density; or if both space and matter be altered, the matter must 

 be changed in the proportion compounded of the two, in which both space and 

 density are varied. Now the space which we suppose the light to occupy, as it 

 is emerging from the surface of the sun, must not exceed the space contained 

 between the spherical surfaces egh,fhl. For that space, as has been observed, 

 is the maximum. This space may be diminished; but if it be diminished, the 

 matter being diminished as the space and as the density jointly, must be more 

 diminished than in the simple proportion of the density. Therefore the diminu- 

 tion of the matter will be the least possible, if, the space being supposed to con- 

 tinue at its maximum, the matter be diminished in the simple proportion of the 

 maximum of the density to the density required. Various formations of the par- 

 ticles of light might be thought of, which would answer this purpose. It might 

 be answered indeed, on the spherical hypothesis, by diminishing the number of 

 spherules, retaining the maximum of their size ; or retaining the maximum both 

 of size and number, if it can be thought reasonable to diminish the density of 

 each particle. But, on any hypothesis, the diminution of matter cannot be less 

 than has been said. That is, the matter of each emission cannot possibly exceed 

 -j-o-Vt °f the maximum. For the proportion of the maximum of the density to 

 the density required, will be found by computation, to be that of 2084 to J , 



VOL, XIII. L 



