1 64 Mr. hersciiel’s Dcjcription of 
fome refpeCt, left at an uncertainty, and our meafures taken at 
different times, and with different degrees of attention, will 
vary on that account. Nor can we come at the true diftance 
of the centers of any two flars, one from another, unlefs we- 
could tell what to allow for the femi-diameters of the flats 
themfelves ; for different flars have different apparent diame- 
ters, which, with a power of 227, may differ from each other 
(as I have experienced) as far as two feconds. 
The next imperfection is that which arifes from a defleTioi) 
of light upon the wires when they approach very near to each 
-other ; for if this be owing to a power of repuliion lodged at 
the furface, it iseafy to underfland, that fuch powers muff in- 
terfere with each other, and give the meafures larger in pro- 
portion than they would have been if the repulfive power of 
one* wire had not been oppofed by a contrary power of the 
other wire. 
Another very confiderable imperfection of thefe micrometers 
is a continual uncertainty of the real zero. I have found, that 
the leaf! alteration in the fituation and quantity of light will 
affect the zero, and that a change in the pofition of the wires, 
when the light and other circumftances remain unaltered, will 
alfo produce a difference. To obviate this difficulty, when- 
ever I took a meafure that required the utmoft accuracy, my 
zero was always taken immediately after, while the apparatus 
remained in the fame fituation it was in when the meafure was 
taken ; but this enhances the difficulty becaufe it introduces an 
additional obfervation. 
The next imperfeChon, which is none of the fmaileff, is 
that every micrometer that has hitherto been in ufe requires 
either a fcrew or a divided bar and pinion to meafure the dis- 
tance of the wires or divided image. Thole who are ac- 
quainted 
