324 



SCIENCE. 



[N. S. Vol. XVII. No. 426. 



where except at the corners of the squares, 

 where two lines cross and form a point. 

 But when we consider that the above dis- 

 cordances involve the errors of both meas- 

 urements, they do not appear unduly large. 

 Divided by 1/ 2, they give for the probable 

 error of a measurement by one method 

 only =h 0".08 for Paris, and ± 0".16 for 

 Helsingfors; and there is no evidence of 

 a systematic arrangement of signs in the 

 differences between the two methods. We 

 may conclude, therefore, that plates meas- 

 ured by the reseau method and without it 

 give identical results within a very narrow 

 margin; nor does irregular distortion of 

 the film appear to have affected appre- 

 ciably the measures made without the 

 reseau. 



The Spectra of Stars of Secchi's Fourth 

 Type: George B. Hale, Ferdinand 

 Bllerman and J. A. Parkhurst. 

 In his early surveys of stellar spectra, 

 Secchi divided the red stars into two great 

 classes (his third and fourth types), whose 

 spectra differ very markedly in their gen- 

 eral characteristics. Subsequent investi- 

 gations by Vogel and Duner confirmed 

 Secchi 's conclusion with regard to the pres- 

 ence of carbon bands in the spectra of stars 

 of the fourth type, but in view of the in- 

 strumental means employed it was impos- 

 sible for these investigators to distinguish 

 the individual lines in the spectra. An 

 investigation of these stars was accordingly 

 undertaken with a three-prism spectro- 

 graph, used in conjunction with the forty- 

 inch refractor and the two-foot reflector 

 of the Yerkes Observatory. Some 250 

 photographs, ranging in exposure-time 

 from a few minutes up to twenty-five hours, 

 were made. They include the yellow and 

 green as well as the blue regions of the 

 spectra. A special study has been made of 

 eight stars, in whose spectra the wave- 



lengths of several hundred bright and 

 dark lines have been measured. The pres- 

 ence of bright lines, though suspected by 

 Secchi, was denied by subsequent observers, 

 but has been abundantly confirmed by the 

 present photographs. Hitherto it has not 

 been possible to identify these lines. A 

 large part of the dark lines, however, have 

 been found to be due to iron, titanium and 

 various other substances. By the aid of 

 these lines the radial velocities of the 

 eight stars have been determined. The 

 photographs bring out a marked resem- 

 blance between the spectra of the two 

 classes of red stars, so far as the dark lines 

 are concerned. Cyanogen is present in 

 both classes, but carbon, either alone or in 

 combination with oxygen, is absent from 

 stars of Secchi's third type, while very 

 conspicuous in stars of the fourth type. 

 In both classes of stars the relative inten- 

 sities of the dark lines in the spectrum of 

 a given element seem to differ considerably 

 from the corresponding intensities in the 

 solar spectrum. This led to a comparison 

 of the stellar lines with the widened lines 

 in the spectra of sun-spots. So far as can 

 be judged from the present photographs, 

 there is a marked similarity of these 

 spectra, but this can not be made the basis 

 of any theoretical conclusions before fur- 

 ther investigations with higher dispersion 

 have been made. In general, the investi- 

 gation tends to confirm the opinion of 

 Vogel and Duner that the two classes of red 

 stars have developed from solar stars. Full 

 details of the work, with tables of wave- 

 lengths and reproductions of photographs, 

 will appear soon in the Publications of the 

 Yerkes Observatory. 



Badial Velocities of Twenty Stars having 

 Spectra of the Orion Type: Edwin B. 

 Frost and "Walter S. Adams. 

 This paper represents a part of the work 



done during the past year with the new 



