ANNIVERSARY ADDRESS. 21 
Ireland on September 1, 2,700 miles to the east. Again, we havea 
record at Maranham, in the north of Brazil, on August 31, 9,000 
miles to the west, so that the dust must have travelled 2,000 miles 
per day westward from Seychelles. An inconsistent observation 
comes from the Gold Coast, September 1, 6,300 miles to the west— 
inconsistent I mean with the supposition that the dust came from 
Krakatoa ; but it is not fatal to that supposition, as a cloudy sky 
may have hidden the first red glow on the Gold Coast. On Sep- 
tember 2 we have the phenomenon recorded at Trinidad, 10,000 
miles west, and at Panama, about 1,000 miles further west. The 
latter case gives a movement of 1,000 miles per day westward 
from Brazil. On September 2 it appeared also in Peru. On 
September 3 we have a record from a ship 4,000 miles west of 
Panama ; but this can scarcely be connected with the westerly 
movement, and we must turn again to the easterly. On Septem- 
ber 1 we had the phenomenon at New Ireland. From that to 
the place of the ship above indicated is an eastward course of 
3,600 miles, giving 1,800 miles per day. Honolulu lies about 
900 miles north-west from the position indicated, or 6,000 E.N.E, 
from Krakatoa, and there the phenomenon occurred first on 
September 5, quite inconsistent with the previous case. But, 
waiving this anomaly for the present, it seems to me quite im- 
possible that upper currents of the atmosphere could carry dust 
from Krakatoa eastward and westward simultaneously at such a 
prodigious rate. If the dust came from Krakatoa, then the only 
agency that seems capable of dispersing it is electrical repulsion, 
and to that view there are serious objections. 
On September 8th we have the first record at Ceylon, 1,600 
miles to the north-west, although at Karachi, 1,400 miles further 
in the same direction, the record is August 28. Ongole, Septem- 
ber 9th, 500 miles northwards from Ceylon, and Madras, not so 
far north, both agree well with the Ceylon observation. At these 
places, as at some others, the first appearance noted was the green 
colour of the sun ; but as this seems undoubtedly to be caused by 
the same agency that produces the red glow, it is not necessary to 
distinguish the two phenomena in giving dates. 
