Mr. herschel’s Catalogue of Double Stars . 113 
the colour a little, making a white flar appear pale red when 
the altitude is not fufficient to clear it of the vapours. It is 
difficult to find a criterion of the colours of (lars, though I 
might in general obferve that Aldebaran appears red, Lyra 
white, and fo on ; but when I call the liars garnet, red, pale 
red, pale rofe-colour, white inclining to red, white, wffite 
inclining to blue, blueiffi white, blue, greeniffi, green, dufky, 
I wiffi rather to refer to the double liars themfelves to explain 
what is meant bv tliofe termsr 
IV. The dillances of the liars are given feveral different 
ways. Thofe that are eflimated by the diameter can hardly be 
liable to an error of fo much as one quarter of a fecond ; but 
here mull be remembered what I have before remarked on the 
comparative appearance of the diameters of Hars in different in- 
(Iruments. Thole that are meafured bv the micrometer, I fear, 
may be liable to an error of almofl a whole fecond ; and if not 
meafured with the utmofl care, to near 2''. This is, however, 
to be underllood only of (ingle meafures ; for the diflance of 
many of them that have been meafured very often in the courfe 
of two years obfervations can hardly differ fo much as half a 
fecond from truth, when a proper mean of all the meafures is 
taken. As I always make the wires of my micrometer outward 
tangents to the. apparent diameter of the (lars, all the mea- 
fures mull be underllood to include both their diameters ; fo 
that we are to dedufl the tw 7 o fern i- diameters of the (lars if we 
w r ould have the diflance of their centers. What I have faid con- 
cerns only the wire micrometers, for my lafl new micrometer 
is of a fuch a conflrudlion, that it immediately gives the diflance 
of the centers and its meafures (as far as in a few months lhave 
been able to find out) may be relied on to about one-tenth of a 
fecond, when a mean of three obfervations is taken. When I have 
Vol. LXXII. added 
