206 Hastings — Optical Constants of the Human Eye. 



The last column beaded 1/A is the well known Donders con- 

 stant of accommodation. This should be nearly constant for 

 all wave-lengths of light. Neglecting the last value, which is 

 sure to be greatly in excess on account of unavoidable errors 

 due to fluorescence of media of the eye impairing the pre- 

 cision of vision in this region of the spectrum, the mean value 

 of the accommodation constant is -00825/1 in. With this 

 value of the constant of accommodation and an assumed accu- 

 racy of the values of c, the calculated values of the far and 

 near points are as follows : 



N' F' 



A . 15-71 18-06 



E> 1-2-50 13-94 



F 10-35 11-31 



G 8-70 9-38 



H.K .__ 7-58 8-09 



The errors implied in the observations of the extreme violet 

 light are just such as we should expect from the imperfect 

 vision in this region of the spectrum due to fluorescence. If 

 we assume that the differences for the other wave-lengths are 

 due to errors of observation only, we find that the probable 

 error of a single measurement is ± 0*067 of an inch. As the 

 measurements seem to have aimed at a precision not much 

 greater than 0*1 of an inch, the accordance must be regarded 

 as very satisfactory. 



As to the precision with which c is determined, for it is the 

 variation of this quantity which yields the physical constant 

 desired, we have the following test. It is known that the 

 media of the eye have dispersive powers not differing in kind 

 from that of other colorless media, hence the differential 

 refractive power of the eye ought to be expressible as a linear 

 function of the indices of refraction of such substances, water 

 for example. Thus, from the following data, 



C A -C n - n A 



A -ooooo -ooooo 



D -01640 -00365 



F ._._ '03300 -00788 



Gr -05125 -01135 



H.K -06827 -01424 



I find, 



c A —c = (4-50 ±'08) (n — ?i A ) 1 /l inch. 



In this result the observation on extreme violet light is 

 included. The general conclusion, when the centimeter is 

 chosen as the unit of length, may be written 



$c/8n = 1-77 /l cm 

 where Bn indicates the variation of index of refraction of water. 



