of Chromatic Dispersion. 367 
in the B wave on entering such a medium will appear to be only 
950,000 instead of 1,000,000, while in the H wave the number 
will appear to be only 520,655 instead of 570,655; so that each 
of these waves has had its length curtailed by the space which 
50,000 zthereal particles would occupy in this medium, in ad- 
dition to the curtailment, in the proportion of 0°66 to 1, due to 
the increased density of the «ther. Now this must be owing to 
the circumstance that the molecules of the medium exert on the 
ethereal particles in its pores an influence such as to increase 
their tendency to remain in their normal positions, to a certain 
extent beyond the measure in which that tendency is increased, 
by the mere approximation of the «thereal particles themselves. 
By the operation of this force, the B wave has its length dimi- 
nished in the proportion of 0°95 to 1, but the H wave in the 
proportion of nearly 0°91238 to 1; so that in the case of the B 
wave the Joss is in the proportion of 0°05 to 1, but in the H wave 
it is 0'08762 to 1, these losses standing to each other in inverse 
proportion to the initial moving forces; and the like may be 
proved with respect to all the other wave-lengths. Thus the 
loss of length arising from that peculiar action of the ponderable 
molecules of the medium on the ether in its pores, in virtue of 
which it increases the persistence of the ethereal particles in 
their normal positions, is in exact inverse proportion to the 
amount of the primary force by which the undulation is excited ; 
and this is precisely what ought to be expected. The absolute 
loss of length sustained by each wave from this cause is the 
same, but its re/ative loss is in inverse proportion to its primary 
length. Thus dispersion arises from the operation of a force in 
virtue of which the ponderable molecules of the medium com- 
municate to the ethereal particles, which are for the instant 
associated with them, a certain amount of persistence in their 
normal positions beyond what is due to the increased proximity 
of the particles themselves. It is constant for the medium and 
temperature ; but it is specific, and independent of the size of 
the pores of the medium in which the ether is momentarily 
more or less compressed; so that a high dispersive power may 
consist with a low refractive power, or vice versa; although in 
general the greater the compressive power of the medium, the 
greater the amount of this peculiar force. 
This, then, accounts for the larger proportion of the variable 
quantity which remains after deducting from the indices corre- 
sponding to the fixed lines the constant which represents the 
density of the «ether in the pores of the medium. That variable 
portion corresponds to a loss of wave-length inversely propor- 
tional to the primary wave-length, and may be found by multi- 
plying the index by this loss, as already pointed out in the case 
