FEBSrAEY 14, 1902.] 



SCIENCE. 



257 



31. 'A Short and General Method of Determin- 

 ing Orbits from Three Observations': A. O. 

 Leusohneb. [Read by O. Stone.] 



32. 'Elements of Asteroid 1900 G A and Ephe- 

 meris for the Opposition of 1901-1902': A. 0. 

 Leuschnek and Adelaide M. Hobe. [Read by 0. 

 Stone.] 



33. 'Discovery of Rapid Motion in the Faint 

 Nebula Surrounding Nova Persei': 0. D. Pek- 

 rine. [Read by W. W. Campbell.] 



34. 'A Determination of the Wave Lengths of 

 the More Prominent Nebular Lines': W. H. 

 Weight. [Read by W. W. Campbell.] 



35. 'The Bruce Spectrograph of the Yerkes Ob- 

 servatory': E. B. Fkost. [Read by G. E. Hale.] 



3ti. 'A Remarkable Solar Disturbance': 

 George E. Hale. 



37. 'A Determination of the Cause of the Dis- 

 crepancy between Measures of Spectrograms made 

 with Violet to Left and Violet to Right': H. M. 

 Reese. [Read by W. W. Campbell.] 



38. 'Four New Spectroscopic Binaries with 

 Notes on the General Subject': W. W. Camp- 

 bell. 



39. 'Discovery of 500 Close Double Stars': W. 

 J. Htjssey. [Read by W. W. Campbell.] 



40. 'Discovery of 300 New Double Stars': R. 

 G. Aitkin. [Read by W. W. Campbell.] 



ABSTRACTS OF PAPERS. 



I'he Flash Spectrum, Sumatra Eclipse, May 



18, 1901: S. A. Mitchell. 



The writer, through the courtesy of the 

 director of the Naval Observatory, became 

 a member of the expedition to view the Su- 

 matra eclipse on May 18, 1901, and was 

 stationed at Sawah Loento. Two instru- 

 ments were employed, a camera of 104 

 inches focus to be used in connection with 

 a ccelostat; and a spectroscope consisting 

 of a Rowland flat grating of 15,000 lines 

 having a ruled space of 3^ x 5 inches, and 

 a quartz lens of 3 23/64 inches aperture 

 and 72 inches focal length. Light from the 

 sun reflected by the ccelostat mirror in a 

 horizontal direction, fell on the grating 

 where it was diffracted, and was brought to 

 a focus on the photographic plate by means 

 of the quartz lens. If grating and photo- 

 graphic plate are each perpendicular to the 

 diffracted beam, the spectrum is 'normal.' 



It was arranged to photograph the first 

 order spectrum from I 3,000 to -'. 6,000. 



The weather on the day of the eclipse was 

 extremely disappointing. First contact 

 was observed in a perfectly clear sky, but 

 clouds soon began to gather and were so 

 dense at second contact that the first flash 

 was not observed at all. Toward the mid- 

 dle of totality conditions became a trifle 

 better, so that it was possible to see, through 

 clouds, the corona extending for about half 

 a diameter from the sun. During no time 

 of the 5 min. 41 sec. of totality was an un- 

 clouded view of the corona obtained, but 

 nevertheless, the second flash was seen 

 beautifully. Altogether eight exposures 

 were made, one before and one just after 

 totality for the cusp spectrum, one at first 

 and one at second flash, and four with dif- 

 ferent lengths of exposure during the total 

 phase. The second flash seemed fully ex- 

 posed, and it is to a discussion of this 

 photograph that this paper is devoted. 



The peculiarities of this photograph are : 



1. Normal spectrum. 



2. Great dispersion. 



On the plate the distance from F to H is 

 95.4 mm., and as the spectrum is normal, 

 1 mm., therefore, corresponds to a differ- 

 ence of wave-length of 9.37 tenth-meters, 

 or 1 tenth-meter corresponds to a dis- 

 persion of about 0.1 mm. For some reason, 

 the spectra were not in perfect focus, but 

 in spite of this fact, in view of the great 

 dispersion of the spectrum, measures were 

 made and wave-lengths determined with 

 a high degree of accuracy. The spectrum 

 extends from ). 4,924 to A 3,320, but the 

 focus becomes poor beyond K, and meas- 

 ures were discontinued at -'. 3,835. For 

 the purposes of the present comparison, 

 the region from F to H only was regarded. 

 In this part of the spectrum 363 lines were 

 measured in the flash. An arbitrary scale 

 of intensities was assumed whereby rep- 

 resents the faintest line seen with certainty. 



