1891.] on An Astronomer^ s Worlc in a Modern Observatory. 405 



The photograpli now on the screen shows the interior of the 

 observatory, and this brings me to the description of observations of 

 an entirely different class. In this observatory the roof turns round 

 on wheels, so that any part of the sky can be viewed from the tele- 

 scope. This is so because the instrument in this observatory is 

 intended for purposes which are entirely different from those of a 

 transit circle. The transit circle, as we have seen, is used to 

 determine the absolute positions of the heavenly bodies ; the helio- 

 meter to determine with greater precision than is possible by the 

 absolute method the relative positions of celestial objects. 



To explain my meaning as to absolute and relative positions :- 

 It would, for example, be a matter of very little importance if the 

 absolute latitude of a point on the Eoyal Exchange or the Bank of 

 England were one-tenth of a second of arc (or ten feet) wrong in the 

 maps of the Ordnance Survey of England — that would constitute a 

 small absolute error common to all the buildings on the same map of 

 a part of the city, and common to all the adjoining maps also. Such 

 an error, regarded as an absolute error, would evidently be of no 

 importance if every point on the map had the same absolute error. 

 There is no one who can say at the present moment whether the 

 absolute latitude of the Eoyal Exchange — nay, even of the Koyal 

 Observatory, Greenwich — is known to ten feet. But it would be a very 

 serious thing indeed if the relative positions on the same map were ten 

 feet wrong here and there. For example, if of two points marking a 

 frontage boundary on Cornhill one were correct, the other ten feet in 

 error — what a nice fuss there would be ! what food for lawyers ! what 

 a bad time for the Ordnance Survey Oifice ! Well, it is just the same 

 in astronomy. 



We do not know, we probably never shall know with certainty, the 

 absolute places of even the principal stars to yV^^ ^^ ^ second of arc. 

 But To^^ ^^ ^ second of arc in the measure of some relative position 

 would be fatal. For example, in the measurement of the sun's 

 parallax an error of yV*-'^ ^^ ^ second of arc means an error of 

 1,000,000 miles, in round numbers, in the sun's distance; and 

 it is only when we can be quite certain of our measures of much 

 smaller quantities than y o^h of a second of arc, that we are in a posi- 

 tion to begin seriously the determination of such a problem as that 

 of the distances of the fixed stars. For these problems we must use 

 differential measures, that is measures of the relative positions of two 

 objects. The most perfect instrument for such purposes is the 

 heliometer. 



Lord McLaren has kindly sent from Edinburgh, for the purposes 

 of this lecture, the parts of his heliometer which are necessary to 

 illustrate the principles of the instrument. 



This instrument is the same which I used on Lord Crawford's 

 expedition to Mauritius, in 1874. It was also kindly lent to me by 

 Lord Crawford for an expedition to the Island of Ascension to 

 observe the opposition of Mars, in 1877. In 1879, when I went to 



Vol. XIIL (No. 85 ) 2 e 



