BRinewoop.—On Red Sunsets. 887 
beneath the surface the velocity is less in proportion to that depth ; 
in like manner, if we look on the atmosphere as part and parcel of 
the earth, at a certain height the velocity is greater in proportion to the 
height. 
The whole world has been greatly interested during the last seven or 
eight months by the beautiful phenomena of coloured suns and brilliant 
sunsets, and the liveliest interest has been exhibited as to their origin. 
Lockyer was the first, I believe, to point out the fact of the phenomenon of 
coloured suns appearing first in the east and then gradually shifting to the 
west. He traces them to Panama, and then speaks of them as having been 
seen on a north and south line; but it strikes me that after leaving Panama 
the phenomenon passed still further westward, seen on the 8rd September 
4,000 miles west of Panama, and at Honolulu on the 5th, and struck India 
and Ceylon on the 8th September, thus performing more than a complete 
circuit of the globe ; moreover, I am of opinion that it may be traced still 
further westward, where it was seen in lat. 24° 06’ N., long. 140° 29’ W., 
by Captain Penhallow of the barque “ Hope,” on the 25th September, 
having then performed 23 revolutions of the globe. 
All the information that I have collected, and from which I have com- 
piled the following tables, has been obtained from ** Nature.” The time 
column has been deduced from the time and date of the phenomena appear- 
ing at the different stations, reduced to Krakatoa time. In some instances 
great difficulty has been experienced, especially in reference to the time at 
Maranham in Brazil, and at Trinidad, and it has been concluded that at 
those two stations the times are late, because it was seen at Panama before 
the time given at them, which we suppose to be an error. Likewise 
in the case of the Gold Coast, in one place the date given is the 30th 
August, and in another the 1st September, but from the general result 
it would appear that it reached that locality about midnight 80th-31st 
gust. 
The tables, I trust, are sufficiently clear. The first column of miles 
represents the mean diurnal velocity that the cloud travelled at between 
Krakatoa and the different localities en route; and in the subsequent 
columns are given the same from each station in rotation. Of course it 
will be understood that a small error of an hour or two in the time at the 
stations comparatively close to the eruption would make a large difference 
were we to show the diurnal velocity: and as I have had only a week’s 
notice to prepare this paper, I trust that any errors that may be hereafter 
found will be treated with that consideration. 
