38 
latter known especially as the ‘ low sun-band.’ 
‘The same result has been obtained in England 
and in America. A careful examination of the 
spectrum with a powerful grating spectroscope, 
made at sunset on Dec. 28, showed that the 
aqueous lines were feeble; and the spectrum, 
at its disappearance, was much farther extended 
towards the green than is usual in a clear sky. 
From all these considerations, it seems that 
the hypothesis of an excess of aqueous vapor 
in the atmosphere is not tenable. 
It seems not unreasonable to suppose that 
the upper regions of the atmosphere have re- 
ceived from some source an accession of light 
matter which reflects the sunlight. Of the two 
suggested sources, — meteoric dust encountered 
by the earth in its progress, and volcanic mat- 
ter projected to an enormous height, — either 
would be a satisfactory explanation. The for- 
mer would seem in itself the more reasonable, 
were there not in this instance special consid- 
erations which give additional weight to the 
latter. Both of these hypotheses have been 
independently suggested by various writers. 
Mr. Ranyard advocates the meteoric view in 
Knowledge for Dec. 7, and Mr. Lockyer the 
volcanic theory in the London Mail of Dec. 
10, and current numbers of Nature. English 
scientific men have shown great interest in this 
investigation; but few references to it have 
been made, as yet, in the publications of other 
countries. 
It will be of interest to classify the dates at 
which the atmospheric phenomenon has been 
earliest observed in different countries. The 
following table contains a list of the dates and 
countries, with the approximate distance and 
direction of each country from the Straits of 
Sunda, in which occurred the tremendous vol- 
canic outburst of Aug. 26. It should be 
noted, that, while the dates given have been 
collated from the best evidence at hand, there 
is a possibility that they may be too late in 
some cases, either from the fact that earlier 
observations have not been reported, or were 
not made owing to unfavorable weather: they 
must therefore be taken as only approximately 
accurate. A few have been derived from gen- 
eral statements in which the exact dates were 
not mentioned. 
This table has been derived mainly from 
English periodicals and from the records 
of the U.S. signal-service. The important 
references to New Ireland and the Hawaiian 
Islands were received by letter from Mr. S. E. 
Bishop of Honolulu, who has also obtained from 
shipmasters the information that the phenom- 
enon has been extensively seen on the Pacific 
SCIENCE. 
ok & 
[Vor. IIL, No. 49. 
Ocean since Sept. 1. It is also reported from 
China, but no date is assigned. 
Distance and direc- + 
Date. Country. tion from Straits 
| of Sunda. 
1883 Miles. 
Aug. 28 .| Rodrigues 3,000 S.W. 
28 . | Mauritius 3,500 S.W 
28 .| Seychelles 3,500 W. 
30 .1| Brazil. 5 10,500 W. 
Sept. 1 . | Gold Coast . 7,500 W. 
. | NewlIreland . ' 8,000 E. 
2 .{| Venezuela . 12,000 W. 
2 West Indies 12,000 W. 
2 Peru) P3.s% ove 18,000 W. 
5 .| Hawaiian Islands. 7,000 N.E. 
8 . {| Southern India 2,000 N.W. 
8. ce WCeylons i330 see eo 2,000 N.W. 
15 Southern Australia . 3,000 S.E. 
15 Tasmania fg Sante 4,000 S.E. 
20 Cape of Good Hope. 6,000 S.W. 
Oct: 8 ~. |) Mlorida ei ee 13,000 N.W. 
19 .{| California ah Se 9,500 N.E. 
20 .} Southern United States 11,000 N.E. 
Nov. 9 . | England . set fukn! 7,500 N.W. 
20) <5.) oinke yrs. clk 7,000 N.W. 
21 +. United States . 11,000 N.H. 
25 .| Italy 7,000 N.W. 
26 .| France 7,500 N.W. 
28 . | Germany 7,000 N.W. 
30 .| Spain . 8,000 N.W. 
30 .| Sweden 7,500 N.W. 
An examination of this table shows at once 
the wide-spread character of the phenomenon, 
and its progressive motion. It is impossible — 
not to conjecture a connection with the vol- 
canic eruption in the Sunda Straits, by which, 
on Aug. 26, the island of Krakatoa disappeared 
wholly from the face of the earth. The terri- 
ble nature of this outburst can hardly be real- 
ized: the sky was darkened for several days, 
the noise was heard two thousand miles, mag- 
netic disturbances were noted, the tidal wave 
was distinctly felt at San Francisco, and the 
atmospheric disturbance was sufficient to cause 
marked barometric fluctuations, which were 
noted by the barographs on the continent, in 
England and America, for several succeeding 
days. Coincidence in dates is not a proof of 
a connection between the atmospheric and the 
volcanic phenomena; but it is certain that the 
former were first observed near the scene of 
the latter, and that similar atmospheric effects 
have been heretofore recorded over limited 
areas in connection with volcanic outbursts. 
Assuming the origin of the atmospheric effects 
to be the volcanic eruption, the table shows an 
extremely rapid progression in both an easterly 
and a westerly direction, —the former over the 
Pacific Ocean, the latter over the Indian and 
Atlantic oceans, to South America and the 
West Indies. Mr. Lockyer considers that the 
latter continued westward to the Hawaiian 
Islands, and does not regard an eastward pro- 
