J stronomical Society. 531 



2d. The Latin translation made bj^ Trapezuntius (George of Tre- 

 bizond), and published by Gauricus, also at Venice, in 1528. This 

 trnslation, from which most of the subsequent editions seem to be 

 copied, is said to have been made from a copy of a Greek manuscript 

 which the Abbe Laurentius Bartolinus caused to be made from one 

 in the Vatican library. 



3d. The Greek edition published at Basil, in 1538, by Grynseus. 

 It is said that he made use of a manuscriirt that belonged to Regio- 

 montanus, who had it from the Cardinal Bessarion, and deposited it 

 in the library of Nuremberg. This, however, has been since doubted ; 

 and it is not quite certain where the manuscript above mentioned is 

 now to be found. This is the more to be regretted, as it is the only 

 Greek edition extant, except the recent one which I am now about 

 to mention. It is evidently not the manuscript from which either of 

 the preceding translations was made. The work is dedicated to our 

 Henry VIII. 



4th. The two other sources are comprised in the Greek edition 

 (accompanied by a French translation), published at Paris, in 1813, 

 by M. Halma. In editing that part of the volume which contains 

 the catalogue of stars, M. Halma availed himself of two additional 

 manuscripts which were in the public library of Paris : one of these 

 he calls the Paris manuscript, which is made the ground-work of his 

 publication, and the other the Florence manuscript. But some other 

 sources must have been appealed to, as he occasionally inserts read- 

 ings which are not to be found in either of these manuscripts, or in 

 the Basil edition above mentioned. 



These several sources of information are all that the public press 

 affords us in our inquiries relative to this interesting subject ; but 

 they are lamentably deficient for the purpose. And, as it is pro- 

 bable that the public are not fully aware either of the amount or the 

 frequency of discordance that exists in this matter, I trust I shall not 

 encroach too much on the patience of the meeting by stating, as 

 briefly as possible, the result of my own investigations and researches. 

 Having carefully compared the position of every star, as given in each 

 of the several copies above alluded to, I have found that out of 1028 

 stars, of which the catalogue consists, there are about 780 (or more 

 than three-fourths) of them that are discordant, either in longitude 

 or latitude; and tliis, not merely in 10, 20, or 30 minutes, which is 

 no uncommon difference, but sometimes to an amount involving 

 whole degrees. That these errors have arisen mostly from the care- 

 lessness of the copyists, and that they may be partly corrected, in 

 some cases, by a reference to the true position of the star, I am ready 

 to admit ; but still there are numerous cases where this tentative 

 method will not avail us, and where we are, after all, left in doubt 

 as to the identity of the star, or the true reading of the original 

 numbers. 



One source of error I have discovered to l)e very common, and in 

 the correction of which I have found the translation from the Arabic 

 to be of essential .service. It is this. The (ireck notation for mi- 

 nutes being denoted by some fractional part of the degree, and such 

 2 M 2 



