292 



SCIENCE 



[N. S. Vol. XXVI. No. 661 



total probable error of a coordinate, made up 

 of errors from all sources, being only 0."39. 

 The nimiber of stars measured in Volume I. 

 is 65,750, and in Volume 11., 66,718. Many 

 of these are duplications, owing to the over- 

 lapping of the plates, but aside from this 

 desirable duplication, the number of different 

 stars measured in the whole Oxford section 

 will be very great. 



VARIABLE STABS 



Two contributions to the subject of variable 

 stars have recently appeared in the Annals of 

 the Harvard College Observatory. 



Volume XLVII., Part I., gives a detailed 

 photographic study, by Mrs. W. P. Fleming, 

 of the comparison stars for the 222 variables 

 of long period, nearly all of which were dis- 

 covered by her by means of their spectral 

 peculiarities. No star having a spectrum of 

 the class designated Md, that of the third 

 type with the hydrogen lines bright, has yet 

 been found which is not variable, although 

 many stars having a different spectrum are 

 also variables of long period. The present 

 volume deals only with the identification, 

 positions and magnitudes of the comparison 

 stars. Rectangular coordinates, referred to 

 the variable star as a center, are employed. 

 The methods of measurement and reduction 

 have been already explained in the Annals. 

 At the time this work was undertaken no good 

 method existed for the determination of pho- 

 tographic magnitudes. Those used in the 

 present volume were derived from measure- 

 ments of the stars with a scale having a series 

 of images of different intensities. Each 

 image, after the first, which had an exposure 

 of one second, had an exposure three times 

 as long as that of the preceding image. The 

 intervals thus obtained are assumed to repre- 

 sent one magnitude. The magnitudes thus 

 obtained are made to depend upon the visual 

 magnitudes of the brighter stars of the se- 

 quence. This method furnishes a scale of 

 magnitudes which, however large systematic 

 errors it may contain, appears to be consistent 

 within itself. Later, when the method which 

 Professor Pickering has devised, or any other 

 method for the determination of absolute pho- 



tographic magnitudes, is available, systematic 

 corrections can be applied to the values here 

 given to reduce them to an absolute scale of 

 magnitudes. A later volume will furnish a 

 discussion of the observations of the variables 

 themselves. 



Volume LV., Part I., contains the Second 

 Catalogue of Variable Stars, prepared by Miss 

 A. J. Cannon. As explained by Miss Cannon, 

 the history of variable star catalogues extends 

 as far back as 1844, when a list of 18 vari- 

 ables appeared, compiled by Argelander. In 

 successive lists the number of objects has in- 

 creased since that time, at first slowly, but 

 later with great rapidity through the intro- 

 duction of photographic methods, until we 

 come to the present catalogue, which contains 

 1,957 variables. This includes the 500 vari- 

 ables found in the globular clusters, but not 

 the 1,800 found in the Magellanic clouds. 

 Altogether, at the present time, about 3,750 

 stars are known to be variable, of which about 

 2,900 have been found at the Harvard Ob- 

 servatory. A study of the number and dis- 

 tribution of the variable stars over the whole 

 sky seems now to be within reach. The pres- 

 ent catalogue is the result of about ten years 

 of compilation and observation. The founda- 

 tion of a card catalogue of variable star litera- 

 ture was begun in 1897, by Professor W. M. 

 Reed. This bibliography, carried forward by 

 Miss Cannon, now consists of more than 35,- 

 000 cards. This vast amount of material, as 

 well as much unpublished data belonging to 

 the Harvard Observatory, has been used' by 

 Miss Cannon in making up the present vol- 

 ume. The main table gives, after the various 

 designations of the star, and its position, the 

 maximum and minimum magnitudes, the 

 period when known, the epoch, class of vari- 

 able, type of spectrum, provisional number in 

 order of discovery assigned by Kreutz, and the 

 date and name of the discoverer. Auxiliary 

 tables and remarks give much information in 

 regard to the peculiarities of many of the 

 variables. Part II. of the same volume will 

 contain further information, including a study 

 of all the published maxima and minima of 

 variables of long period. S. I. Bailey. 



