176 Mr. M. Ponton on certain Laws 
tion to its primary length, from the operation of the dispersive 
power of the medium alone; and these are in inverse proportion 
to the primary wave-lengths. Thus also the proportion of the — 
refractive indices corresponding to this temporary loss of wave- 
length, must also have a certain dependence on the initial force 
which generated the particular wave to which the index belongs, 
and may be found by multiplying the indices “B,, “Co, &e. (being 
the observed indices freed from the portion due to the extrusion) 
by the above complementary numbers, representing the loss of 
length sustained by the waves from the operation of the disper- 
sive power. Thus, in the case of the bisulphuret of carbon, the 
indices “B,, “C,, &c. are 
HB, 1-615760, “C,1-620667, “D, 1632958, HE, 1-647542, HF, 1°659425, 
G,1-679311, “H, 1695681. 
These, multiplied by the above series of complementary numbers, 
give for the proportion of the indices due to the dispersive power, 
B0-095255, C0:100162, D0:112453, E0:127037, F0-138920, G0-158806, 
H0175176, 
which numbers are identical with the differences between the 
above indices and the constant e=1°520505. 
In different media, the loss of length sustained by any one 
wave through the action of the dispersive power is always pro- 
portional to the constant a, which may be accordingly regarded 
as a measure of that loss. 
To generalization beyond this pomt, the uncompensated errors 
of observation and the yet unascertained effects of change of tem- 
perature are a serious obstacle. But the data already obtained 
may be found useful in detecting some of those errors and effects, 
and in determining their probable limits. 
As regards the effects of temperature, the most instructive 
cases are those of the oils of cassia and anise; for of these we 
have observations at three different temperatures, though unfor- 
tunately these do not coincide in the two media. Comparing 
the values of ¢ in these two fluids for the three sets of observa- 
tions, they will be found as under :— 
Oil of Cassia. Diff. Oil of Anise. Diff. 
No. 1. Temp. 10° ¢=1:477740 . 2478 | No. 1. Temp. 13°2e=1-478492 . 989 
pee es ae 1475262 . 5726 2.5 - 15% > 1°4775038 20a 
3. , 22°5 1:469536 . 8204 3. 5 20°9 1:4738459 . 5083 
It will be perceived that not only are these values of ¢ in the 
inverse order of the temperatures, but that their differences are 
not far from being proportional to the differences of temperature. 
To make them exactly so, they would have to be altered thus :— 
