FOR CORRECTING OBJECT CLASSES. 23 



is termed " outward coma," and indicates that the front incident 

 surface of the back triple is too convex. If, on the other hand, 

 the bright spot is on the outer side of the rings, or next the margin 

 of the field of view, there is " inward coma," which shows that this 

 same surface is too flat. I have previously remarked that this 

 curve has a powerful effect on the flatness of field, and perfection 

 of the oblique pencils, and for them no other correction is generally 

 requisite than an alteration in this radius. 



Before the glasses are finally cemented in their cells, they should 

 be carefully tested for centering ; for this purpose a very minute 

 globule is selected and placed exactly in the centre of the field. If 

 the bright spot appears excentric, the pair of lenses which occasion 

 the error should be shifted on each other while warm enough to 

 cause the Canada balsam by which they are cemented together to 

 yield, till on repeated trial the error is corrected. This is impor- 

 tant, as the least fault in centering materially impairs the per- 

 formance of an object-glass. 



There is yet one other globule test for object-glasses to indicate 

 accuracy of workmanship, or whether the lenses are worked to 

 true spherical surfaces. If the rings from a minute globule appear 

 of an irregular wavy outline, either approximately to a polygon or 

 a triangle, it shows that one of the surfaces at least that refracts 

 the rays is of this form. Such workmanship is inexcusable, and 

 those that cannot avoid it had better let glass grinding alone. 



Finally, there is an appearance that I have sometimes seen in 

 our best object-glasses, when focussed away from a globule, viz., 

 " Newton's rings." This shows that in the contact surfaces of one 

 of the pair of lenses the convex is deeper than the concave, and 

 bears hard in the middle. This may have no worse effect than the 

 loss of light, but still it is as well avoided. 



