718 PHILOSOPHICAL TRANSACTIONS. [aNNO 17Q6. 



the same constellation 4 m, there could he no doubt of a change in the light of 

 Algol when it was found to be not brighter than ^ . But let us in the next place 

 take an observation recorded in my journal. 



" May \2, 1782. (3 Lyras is much less than y." 



Now, examining the British catalogue, we find |3 3 m, and y 3 m. Had the 

 method of orders been adopted by Flamsteed, we should at once have pronounced 

 this star to be changeable. For it would have been j3y in his time, and y|3 at the 

 time of observation ; but since we liave shown that no inference can be drawn 

 from the order of the letters, we have only the magnitudes to refer to. And 

 here again the deviation of |3 from its usual brightness not being so considerable, 

 but that a star such as it appeared to be at the time of observation might pass 

 for one of the 3d magnitude, we are left in the dark. ; and yet, a few years after, 

 this star was actually found to be not only changeable, but periodical. 



M. de la Lande, in mentioning the change of S Ursse majoris, arranges the 

 7 bright stars of that constellation as they appeared to him ; and remarks that 

 sometimes y and £ should stand before (3, and sometimes after it. Here we have 

 something like an order of 7 remarkable stars ; but as it happens, the stars 

 themselves are not favourable to the formation of a regular series. Mr. Pigott 

 and Mr. Goodericke also compared the stars, whose changes they were examin- 

 ing, to other neighbouring stars that were proper to be estimated with them, and 

 were in a manner forced to lay aside the method of magnitudes. These instances 

 contribute to support the arguments above used, to show that another method 

 of ascertaining the lustre of the stars is required, while at the same time they 

 sufficiently indicate that the comparative brightness of stars is the only safe one 

 to which we can have recourse. 



It will be necessary now to enter into a full display of my proposed method ; 

 for simple as it is in its principle, it is not only difficult but very laborious in its 

 progress. I began to put it in execution about 14 years ago ; but other very 

 interesting astronomical pursuits have broken in upon the regular continuation 

 of it. By relating the difficulties or inconveniences as they happened, it will 

 appear that my present notation, as well as method of arranging the observations, 

 are liable to the fewest objections. The general disposition of the stars is in 

 constellations. This order is to be preferred to that of right ascension, or polar 

 distance, because the stars being to be compared to the nearest proper stars that 

 can be found, the constellations themselves will generally answer that purpose 

 better than other selections. 



My first design was to draw each whole constellation into one series. Ac- 

 cordingly I began July l6, 1781, to arrange the stars in Ophiuchus thus: " Or- 

 der of the stars in Ophiuchus; ix(iSl^y]y.yi." This way of placing the stars agrees 

 so far with my present one, that any star, such as k for instance, may be taken. 



