In the year 1874, the late M. Schjellerup, the eminent Danish astronomer, published a French translation of two Arabic manuscripts written by Al-Sufi and entitled "A Description of the Fixed Stars." One of these manuscripts is preserved in the Royal Library at Copenhagen, and the other in the Imperial Library at St. Petersburgh.[390] Al-Sufi seems to have been a most careful and accurate observer, and although, as a rule, his estimates of the relative brightness of stars are in fairly good agreement with modern estimates and photometric measures, there are many remarkable and interesting differences. Al-Sufi's observations have an important bearing on the supposed "secular variation" of the stars; that is, the slow variation in light which may have occurred in the course of ages in certain stars, apart from the periodical variation which is known to occur in the so-called variable stars. More than 900 years have now elapsed since the date of Al-Sufi's observations (about A.D. 964) and over 2000 years in the case of Hipparchus, and although these periods are of course very short in the life-history of any star, still _some_ changes may possibly have taken place in the brightness of some of them. There are several cases in which a star seems to have diminished in light since Al-Sufi's time. This change seems to have certainly occurred in the case of θ Eridani, β Leonis, ζ Piscis Australis, and some others. On the other hand, some stars seem to have certainly increased in brightness, and the bearing of these changes on the question of "stellar evolution" will be obvious. In most cases Al-Sufi merely mentions the magnitude which he estimated a star to be; such as "third magnitude," "fourth," "small third magnitude," "large fourth," etc. In some cases, however, he directly states that a certain star is a little brighter than another star near it. Such cases--unfortunately not numerous--are very valuable for comparison with modern estimates and measures, when variation is suspected in the light of a star. The estimates of Argelander, Heis, and Houzeau are based on the same scale as that used by Ptolemy and Al-Sufi. Al-Sufi's estimates are given in thirds of a magnitude. Thus, "small third magnitude" means 3⅓, or 3·33 magnitude in modern measures; "large fourth," 3⅔ or 3·66 magnitude. These correspond with the estimates of magnitude given by Argelander, Heis, and Houzeau in their catalogues of stars visible to the naked eye, and so the estimates can be directly compared.