THE MICROSCOPE. 71 



2. If a person with a spider line micrometer wished to com- 

 pare two ruled plates under a hi^h power, he is liable to an error 

 which was entirely avoided by the method used above. The lines 

 will usually be covered with a glass (Fasoldt's are on thin glass 

 which is inverted on the slide), and if the cover-correction on the 

 objective is used, the magnifying power will be changed quite sen- 

 sibly; so unless the covers are nearly of the same thickness errors 

 will arise that will vitiate the results: similar trouble may of course 

 arise in any measures with a spider line or other eye-piece micro- 

 meter, as in measuring blood-corpuscles, etc. Though this point is 

 probably not new, it is not generally noticed by practical workers, 

 and so some measures may be given here. 



With a Gundlach 4- the circumference of the correction-collar is 

 divided into 10 parts, tenths of a division being easily estimated; it 

 can be turned through 9.6 divisions, or for convenience say 9°. 6. 

 With this objective and a Powell and Lealand spider-line micro- 

 meter (whose screw has a thread of J^^ inch.) Y-ioiy i"ch on a covered 

 Rogers plate equal: 



2.925 turns, collar at o". 



2.98 " " 5''.2 

 3.05 " " 9".6 



2.99 " " 5". 2 



The correction was made by focussing first on the line and in 

 the usual way, then by turning the collar bringing dust on the 

 cover into focus, and then the line again with the fine motion screw. 

 Other methods might give somewhat different results. These were 

 quite consistent. Of course the same part of the same two lines 

 was always used. The lines could be seen well whatever the adjust- 

 ment; this was correct only at 5°. 2: evidently by turning it from 0° to 

 9°. 6 the lenses have been brought so much nearer together as to in- 

 crease the magnifying power by one-twentyfifth. 



A dozen years ago Cross measured the increase of focal length 

 caused by this change, and by his method I find with: 



Collar at o" f == 0.198 inch. 



" 9". 6 = 0.1915 to 1S95 " 



a shortening of about i part in 25: the lens evidently magnifies a 

 very little more than a true -\ would do. 



