146 TESTING THE MICROSCOPE. 



in the case of the low powers, which are usually employed only 

 to obtain a larger field, the degree of perfection is of less im- 

 portance, and consequently the improvement of these powers does- 

 not appear to be so remunerative. We must not, however, 

 estimate too lightly the advantages of a good low-power objective ; 

 on the contrary, we shall learn always to value them highly 

 for all investigations, where sharpness of image and extensive 

 field have to be combined, as, for instance, in researches upon 

 fibro-vascular bundles. 



The testing of Microscopes by means of dioptric images, which 

 are gradually diminished to the limit of discrimination, admits, 

 of course, of various modifications. Instead of wire-gauze we may 

 choose any other objects, spherical, thread-shaped, square, &c., 

 and employ as the standard of comparison not only the power of 

 discrimination, but also the capacity of recognizing form, as, tor 

 instance, the angles of a square or hexagon. Two or more minute 

 perforations in an opaque screen form also a useful test-object, 

 inasmuch as the distance of the air-bubble can be so chosen that 

 the single points in the image are seen just separated from each 

 other. It follows, of course, that the results obtained by any 

 one method are comparable only with each other, and not with 

 the figures found in another way. For the recognition of the 

 form of a single square we shall, with a given amplification, obtain 

 limiting values altogether different from those obtained for the 

 discriminations of the square meshes of a wire-gauze, and the 

 same differences between the absolute magnitudes also appear 

 with the naked eye. On the other hand, it cannot well be 

 supposed that the relative magnitudes, as represented by the 

 curves, should be essentially different. 



We have made use of several of these modifications, and have 

 invariably found that none afford the same certainty of judgment 

 as the one first mentioned. The limit of recognition, especially 

 of spherical and thread-shaped objects, is by no means easy 

 to determine, and is therefore rather variously estimated by 

 different observers. Further, a condition is reached in which we 

 no longer observe the sharp image of the object, but its feeble 

 image formed of dispersion circles, which remains visible some- 

 what longer because of its greater extent. This is evident from 

 the fact that the calculated size of the retinal image with 

 thread-shaped objects is often found decidedly smaller than the 



