8 Mr Barlow On the Practical Construction 
This differs from the former index by 0026, and is given as 
an instance of extreme aberration ; no greater difference than 
this can be allowed ; should it ever exceed this quantity, the ob- 
servation should be repeated. In a great number of such ex- 
periments I have generally found a complete agreement in the 
first three places of decimals : 
The third line of observations gives Index = 1*6013 
The fourth 1*5994 
The first ----- — 1*6019 
The second ------ 1*5993 
4) 6-4019 
Mean Index, 1 *6005 
Similar operations for the plate prism give for a mean in- 
dex r — 1.5279. 
9. Instrument for measuring the Dispersion , and for determi- 
ning the Dispersive Ratio. 
It is a well known optical fact, that light, in passing from one 
medium to another, is not only refracted, but is decomposed in- 
to different coloured rays, thereby forming a spectrum, and that 
the extreme red ray is the least refracted, and the extreme 
violet the most. The indices of refraction for these two ex- 
tremes are therefore different, and the difference between these 
indices divided by the mean index minus 1, is called the Dis- 
persive Ratio ; and the ratio between the dispersive ratio for 
two different species of glass, is called the Ratio of the Disper- 
sive powers, or Ratio of Dispersion . This, also, is sometimes 
called the dispersive ratio of two glasses. 
The instrument for determining this ratio may be described 
as below : 
AB, Fig. 5. is a brass pillar, on the top of which fits the 
cap C, surmounted with a joint K ; to the upper part of which 
is fixed a short tube Imno , open on the side a 5, having a set 
screw s. Within this short tube is inserted another tube of 
about double the length, and which, when brought into any re- 
quired position, may be fixed there by the set screw s shutting 
