40 A. W. Roberts.—Computation of [ Nov. 25, 
“COMPUTATION OF ORBIT OF THE COMES OF SIRIUS. 
By A. W. RoBERTs. 
[READ 25TH NOVEMBER, 1891.] 
I.—INTRODUCTORY. 
‘THE orbit of any body moving in a curvilinear orbit can be 
-determined either by a graphical or by an analytical method. 
In a case of a body whose co-ordinates, either with reference to the 
_earth, or the superior body round which it revoives, can be determined 
«with very great precision, there can be no hesitation as to the method » 
of procedure. 
And yet even as an important auxiliary to the rapid determination 
of the orbit of a comet from early observations the graphical method 
“merits our favour. When later positions are obtained these approxi- 
-mate positions can be corrected by the “variation of curtate dis- 
tances.” | 
With regard to binary stars the graphical method, for many reasons, 
_is to be preferred. 
The latitude of error precludes the more exact refinements of 
analytical research. We are therefore compelled to accept what seems 
the most probable orbit, and the computer strives to obtain an orbit 
that shall best agree with the observed measurements. 
The following is an attempt to construct an orbit for Sirius that 
will not depart much from the observed path. The measures used 
-are those collated by Mr. Gore, and published, along with an orbit, in 
the June number of the Monthly Notices of the Astronomical Society, 
- 1889, 
The orbit computed in this paper differs in one or two particulars 
from the one given by Mr. Gore. In both orbits the periodic time is 
about fifty-eight years. The Comes of Sirius according tothis paper 
is at its minimum distance, fwo seconds, in 1892 ; according to Mr. 
Gore’s orbit in 1892, the two stars are three ana a half seconds 
apart, the peri-astron date being 1896. 
The difference is mainly owing to the much greater eccentricity 
« obtained in this paper. 
.It would be presumptuous to offer an opinion as to the merits of 
