1894.] MICROSCOPICAL JOURNAL. 101 



known, made an average of 1-3224 by measuring from the 

 outside of the dark border on one side to the inside of 

 the dark border on the other. By making the proper cor- 

 rection, so as to include the dark border on both sides, 

 which should be measured the average is 1-3173. Dr. 

 Ewell from a measurement of numerous cases also reaches 

 a similar conclusion. This concensus of opinion in favor 

 of larger averages, is due I think to greater accuracy in 

 the character of instruments employed, and also more 

 particularly to a better knowledge of optics. A human 

 blood corpuscle consists optically of two kinds of lenses, 

 the centre a bi-concave, dissipating the light, and the 

 ring around it, a double convex, concentrating the rays. 

 Outside of this ring is a black border, which was sup- 

 posed to be caused by the diffraction of light striking the 

 edge of the corpuscle. More careful consideration of 

 the subject, however, shows that this black border be- 

 longs to the corpuscle, and that its color is due to the 

 total refraction of light at the edge. There is of course 

 also diffraction, but a wide angle immersion objective 

 which is indispensable for such work, will pick up these 

 scattering rays, and by its accuracy of correction and 

 absence of depth of focus, enable the careful observe* to 

 focus exactly on the edge, where the breadth of the dif- 

 fracted cone is at a minimum. Of the two kinds of eye 

 piece micrometers, I much prefer the filar with both 

 cobwebs adjustable. On a large number of observations, 

 perhaps, the glass eye-piece micrometer would be equally 

 accurate, the errors balancing one another, and it is much 

 easier, more rapid, and less fatiguing to use but the unit 

 of measurement is much larger, with a 1-10 inch immer- 

 sion objective, 12 inch tube, I inch eye-piece, and collar- 

 correction nearly closed, the value of one division of the 

 eye piece micrometer is about 1-80,000 of an inch, which 

 can be estimated to one half or 1-160,000 of an inch, 

 while with a filar micrometei under similar condition^ {?* 2^- 



; « V 



l^j LIBRARY 



