RECORDS 43 



a fluorescent glow similar to that in a Crookes tube. The 

 source of the electrons may be either the highly-heated meteor 

 — in which case the long-continued luminosity of the train 

 must be accounted for by a retardation of the fluorescence pos- 

 sibly due to the low temperature — or the electrons may come 

 from the sun — in which case the explanation would be sim- 

 ilar to that lately given by Arrhenius for the light of the au- 

 rora. The author stated that this last hypothesis has not, to 

 his knowledge, been previously advanced. The balance of evi- 

 dence seemed to show that the luminosity of the persistent 

 trains must be primarily caused by energy of an electrical 

 nature. The subject is one of practical importance owing to 

 its bearings on meteorology. 



The paper by Dr. Mitchell gave the results of observations 

 on the flash spectrum, made by him at Sawah Loanto, Sumatra, 

 during the eclipse of May i8, 1901. Through the courtesy of 

 the Astronomical Director of the Naval Observatory, Dr. 

 Mitchell became a member of the expedition sent out by the 

 government to observe this eclipse. The spectroscope em- 

 ployed was a Rowland objective plane grating of 15,000 lines, 

 used in connection with a coelostat. The weather experienced 

 at Sawah Loanto was like that at almost every other astronom- 

 ical station in Sumatra, cloudy during totality. Through the 

 clouds, nevertheless, a spectrum of the flash at the third con- 

 tact was obtained, which showed 374 bright lines between F 

 and H. Investigations into the reasons for the differences in 

 intensities in the flash and the Fraunhofer spectrum, showed 

 that the intensities depended upon the heights to which the re- 

 versing layers of the different metallic elements around the sun, 

 extend. It was found possible to arrange the elements in three 

 groups according to their atomic weights. 



Comparisons were made with Norman Lockyer's list of "en- 

 hanced " lines, or those stronger in the spark than in the arc, 

 to confirm, if possible, Lockyer's idea that the "enhanced" 

 lines play an important role in the chromosphere spectrum. 

 Fifty-seven per cent, of the " enhanced " lines of titanium were 

 found in the flash, but at the same time all of these lines cor- 



