[ 368 ] 
to A3, by the fmall angle DCF. Let any ftar be 
made to run along the wire A B, from A to B, any 
other ftar following it will pafs the wire D E, at K 
fooner, than the horary circle F G at L, by the time 
of its defcribing the fmall fpace L K, and confe- 
quently the difference of right afcenfion will appear 
too little j now let the wires be turned a quarter 
round, that the wires A B, DE, may change places, 
D coming into the place of A, and E into that of 
B, which is done by making the fir ft ftar run along 
the wire D E, from C to E. Now the wire A B, 
deviating from C I perpendicular to D E, by the angle 
B C I, the fecond ftar will pafs the wire BC at N, 
later than the horary circle I C at M, by the time it 
takes to defcribe the fpace M N — LK; and con- 
fequently the difference of right afcenfion of the two 
ftars will appear as much too great, as it before ap- 
peared too little, when the wires were adjufted in 
their ufual pofition : and half the difference will be 
the corre&ion in this cafe, to be added to the firft, 
or fubtra&ed from the fecond difference of right 
afcenfion ; whence the correction may be eafily in- 
ferred for all other obfervations. 
Celestial 
