95 
We see that the form I have proposed is deducible from 
this by multiplication of each by 
The possibility of obtaining my result was precluded by 
the assumption {n = l) on page 424, The appearance of the 
terms proposed by Professor Rowland is avoided in this 
paper by the condition that the coefficients shall on integ- 
ration tend to a zero limit, if the boundaries are at an 
infinite distance, whereas four of the terms in his coefficients 
will tend to a constant limit. That they disappear on 
integration over the plane (page 435) is perfectly true, but 
as I have argued earlier, this is not the condition required, 
but that no portion of a boundary at an infinite distance, 
even of a small fraction of a wave length, shall produce a 
finite effect (compared with its size) at a finite distance from 
the origin. 
I intend in another paper to discuss the question of the 
direction of the vibration of plane polarised light. 
“On the Different Arrangements of Equal SphericaJ 
Granules, so that the mean density may be a maximum,” 
by Professor OsBOENE Reynolds, F.R.S. 
In a paper on “ The Dilatancy of Granular Media,” read 
before the British Association, at Aberdeen, I pointed out 
that uniform spheres might be so arranged that each 
sphere being held by the adjacent spheres the mean density 
of the mass within the interia was anything between 
— D and — D, 
6 6 
D being the density of a sphere. 
