566 PICKERING— THE OBJECTIVE PRISM. [April 20, 



her. Of the nineteen new stars, known to have appeared during 

 the progress of this work, she discovered ten, and five more were 

 found by other observers here. In this work also, the number of 

 stars of the peculiar class known as fifth type, has been increased 

 from seventeen to one hundred and eight. 



For a more detailed study of the bright stars, prisms have been 

 attached to the ii-inch Draper Telescope at Cambridge, and to the 

 13-inch Boyden Telescope at Arequipa. Spectra of the brightest 

 stars have thus been obtained, six inches long, and half an inch 

 wide, showing at least five hundred lines. Prisms twenty-four 

 inches in diameter have been used with the Bruce Telescope in 

 Arequipa, and sixteen inches in diameter with the Aletcalf Tele- 

 scope in Cambridge. The latest and largest investigation under- 

 taken here, as part of the Henry Draper Memorial, is a catalogue 

 giving the class of spectrum of a hundred thousand stars of the 

 eighth magnitude and brighter, shown on the photograph taken with 

 the 8-inch doublets. The classification of spectra used in the Draper 

 Memorial, has been accepted by the superintendents of the prin- 

 cipal nautical almanacs in their standard catalogue of three thou- 

 sand stars, and also at the leading observatories. The preparation 

 of the catalogue mentioned above has been undertaken by Mrs. 

 Fleming's successor. Miss Annie J. Cannon, who has devoted a 

 large part of her time during the last fifteen years to the detailed 

 study of stellar spectra. Her classification of one thousand stellar 

 spectra published in Volume 28 of the Harvard Observatory 

 Annals, occupied her for three years. To complete, in a reasonable 

 time, a catalogue of one hundred thousand spectra evidently re- 

 quired the most careful study of the methods of " scientific manage- 

 ment." As a first step, her contribution to the work, which re- 

 quired the greatest skill, was reduced from one hundred to six per 

 cent., the remainder being performed as a great piece of routine 

 work, by less experienced assistants. The utmost care has been 

 taken to maintain the highest degree of accuracy, the probable 

 error of the result for each star being about a tenth of one interval, 

 corresponding to four one-hundredths of a stellar magnitude. Miss 

 Cannon is now classifying five thousand spectra a month, and has 



