728 



8CIENCB 



[N. S. Vol. XXVIII. No. 725 



that the moon in the first quarter was un- 

 obscured by clouds most of the time. 



Poet Teert, N. Y., W. E. Ellis 



September 4, 1908 



P. S. — Since furnishing the above, Mr. 

 Donald Eobertson, of Brooklyn, N. T., writes 

 me: 



I saw another illumination last Friday, the 

 fourth of September. I was at Lake Placid in 

 the Adirondacks and had a fine view of it. It 

 began at about eight o'clock and lasted until about 

 nine or ten — I do not know which, as there was 

 always a glow in the sky from eight to ten. In 

 most respects, it resembled the one we saw at 

 Murray Bay, but there was one diflference. The 

 heavens were lit up brighter than on August 18 

 and there were at times rainbow colors to be 

 seen in the north. 



I also wish to add that there was nothing 

 in the displays of August 18 and September 4 

 that remotely resembled the steady eastern- 

 western luminous patches of March 27. I am 

 still of the opinion that the last was something 

 more than an auroral exhibition. 



Since writing my former account I have 

 received several letters from scientists furnish- 

 ing me opinions and latest explanations of 

 auroral phenomena. The explanation that I 

 advanced was not so much at variance with 

 latest views. 



In nearly all of the theories that have come 

 to my notice, it is assumed that the sun is 

 the only source of cathode rays, or the sun's 

 action produces the cathode rays in our own 

 atmosphere. The possibility that the earth 

 may emit its own cathode rays does not ap- 

 pear to have been considered. W. E. Ellis 



FoET Teret, N. Y., 

 September 9, 1908 



To THE Editor of Science : In your issue of 

 July 10 Mr. "Wilmot E. Ellis refers to a re- 

 markable case of illumination of the heavens 

 which was observed in New Jersey on March 

 27 of this year. I was privileged to witness 

 this same phenomenon from the uppermost 

 deck of the R.M.S. Adriatic about 530 knots 

 due east from Sandy Hook Light Ship. I had 

 on several previous occasions observed spots 

 or streaks of small dimensions in various parts 

 of the northern heavens, sometimes rivaling 



in brightness the Milky Way itself, but never 

 had I known of sky-brightness as extensive 

 and as pronounced as I did in this case. The 

 night was perfectly clear and "Venus shone so 

 brightly that a streak of light was thrown 

 upon the surface of the ocean. I watched the 

 phenomenon during the last twenty minutes 

 of its manifestation, the general character- 

 istics being as recorded by other observers, 

 though I believe its general position was lower 

 on the western horizon. I noticed no trem- 

 bling whatever in the shafts of light which 

 developed towards the last. 



I can hardly agree with your correspondent 

 as to the nature of this brightness. His the- 

 ory is practically that proposed by Chaplain 

 Jones, U.S.N., in 1855 for the zodiacal light. 

 I do not pretend to deny a corona to the earth 

 and moon, but if due to disintegration of 

 matter, as Mr. Ellis suggests, it is evident that 

 to allow for an applicable corona we should 

 require a degree of ionization of the atmos- 

 phere close to the earth — apart from that due 

 to sunlight — incomparably greater than what 

 we observe. 



I have long believed that sky-brightness was 

 either due to an after-glow on banks of mois- 

 ture or dust in the upper reaches of the at- 

 mosphere, or else to the burning of cosmic 

 dust clouds as the earth's atmospheric mantle 

 passed through them. These dusts might 

 either be traveling in set orbits like meteorites, 

 or more probably they would emanate from the 

 sun, either as a continuous outgrowing corona, 

 or likely enough as an intermittent discharge 

 from spots. Sunspot areas are, if anything, 

 hotter than the rest of the photosphere, which 

 would be explained by the heat-retaining power 

 of material dusts (probably gaseous in char- 

 acter) present in the direct line of bolometric 

 measurement. The magnetic disturbances 

 often observed shortly after sunspot eruptions 

 would be explained by the bridging of space 

 by these atomic dusts, and their effect as ions 

 or carriers of electricity from the earth to the 

 sun. The time elapsing between the eruption 

 and the magnetic effects would indicate the 

 speed of travel of these particles which might 

 be compared with the growth of solar red- 



