1893.] 



On a Portable Ophthalmometer, 



5 



tion of the constants, especially I and F. I has been determined 

 exactly to 1 /500th inch, and can be estimated to abont l/1000th. 

 The focal length of the object-glass can be determined by Cornn's 

 method, but in general it is more convenient to measure two curved 

 surfaces whose radii are exactly known, and within the limits of the 

 instrument. 



The index error is found by taking the number of dioptres at suffi- 

 ciently great intervals within the limits of the instrument. In this 

 instrument, if we take the extremes of the index, = 12 mm. and 

 = 16 mm., we find the corresponding dioptres are 38*9 D and 

 51 "84 D. The index being graduated in thirds of a millimetre, the 

 index error of each division is nearly 0*08 D, which is positive. 



If the double prism be now removed, the image being single, and 

 the pupillary opening generally distinctly visible, it affords a means 

 of determining whether the visual axis passes through the centre of 

 the pupil. 



It will be seen that this instrument differs from the ophthalmo- 

 meter of Helmholtz, the most perfect instrument theoretically and 

 practically which has- been devised for this purpose, in which, while 

 the object is constant, the image varies with the curvature of the 

 surface, but always covers the same angular interval of the surface. 

 It resembles the ophthalmometre pratique of Javal and Schiotz, in 

 which the doubling is effected by means of a double-image prism in- 

 serted between two achromatic lenses of equal focus, so that while 

 the image is constant the object is made to vary. With this instru- 

 ment, when the difference of curvature of the principal meridians is 

 considerable, amounting to 3 or 4 dioptres, in order to obtain 

 approximately accurate results it is necessary to insert birefractive 

 prisms of different powers, giving images of from 1 to 3 mm. In 

 the present instrument the image of 2 mm. has been selected as giving 

 sufficiently accurate results for most practical purposes, measuring 

 with precision, as it does, a difference of refraction of half a dioptre. 

 For cases outside the limits already referred to (6 to 10 mm.) prisms 

 of suitable powers can be substituted. 



