169 
would still form a fine miniature. It might be reasonably 
expected that such slight errors as had escaped the notice of 
eminent opticians would not materially injure a miniature 
image in which the aberration would probably be reduced in 
the miniature itself. The image of the thermometer now 
formed with the imperfect eighth was viewed with a fine 
sixteenth (at about 800 diameters), when it was gratifying to 
observe a very beautiful display of the picture well defined 
in all respects (fig. 4). 
These and other experiments appeared to warrant an im- 
portant conclusion—that an image-test miniature formed by 
an objective of fair quality enjoyed sufficient accuracy of 
definition in miniature (even when the object was placed at 
varying distances from the stage or focal point of vision) to 
form a trustworthy test of microscopical definition, provided 
the aperture of the miniature-forming objective was equal to 
that of the objective to be tested. © 
To estimate the size of a miniature (a) of a given object (6) 
placed at a considerable distance (d) from the miniature 
(fig. 1 a), it is necessary to consider that the conventional 
focal length F of an objective may be defined to be 10 inches 
divided by the micrometric ratio of amplification (m) when 
the image is thrown on a screen 10 inches from the object, 
so that 
10 
— when d=10, 
or 
2, 
=F— 
co 7, when d is constant ee Boy CN 
m 
When very deep objectives are used (fig. 1) the position of 
the plane of focal vision varies very slightly for a consider- 
able increase of the length of the microscope, so that if the 
draw tube be graduated, the increase of power is nearly 
proportionate to the increased length or reading, because, the 
focal plane being nearly fixed, the image will appear upon a 
screen enlarged proportionally as its distance (d) increases. 
On reversing the rays, the miniature diminishes in propor- 
tion as a given object is removed further from it: from 
which it follows that approximately, and sufficiently near for 
the purpose in hand, 
m cd when F is constant. . . . (2) 
Compounding these expressions (1) and (2), 
pie 8 
To, —P > 9 
