38 



they are caused by any mere meteorological conditions of the air ; 

 conditions which, in the first place, are local, and in the second, 

 are not uncommon. The sunsets, on the other hand, are not local, 

 and they are in a most singular degree uncommon — as uncommon, 

 it might almost be said, as the eruption of Krakatoa itself. The 

 correspondence of time, too, is very well marked : the first pheno- 

 mena of the kind Avere observed in India within a few days of the 

 eruption : a green sun was reported from ]NIadras on the 5th or 6th 

 September; and such appearances have gradually spread to more 

 distant regions. The shock of the eruption appears also to have 

 been transmitted through the air, two or thi-ee times right round 

 the globe ; disturbances following each other at intervals of about 

 thirty-six hours, were shown by every self-recording barometer in 

 Europe ; and though there is not as yet any direct evidence to 

 connect these disturbances, or the extraordinarj"^ sky effects, M'ith 

 this tremendous explosion, the presumption that there is such a 

 connection is very strong.* 



* [A few days later (8th December) a long and detailed letter to the same 

 effect, by Mr. Norman Lockyer, appeared in the "Times" ; and since then, many 

 very valuable communications have been published in "Nature." Two papers 

 on the barometric di.sturbances, by Mr. E. H. Scott and General Strachey, are 

 published in the Proceedings of the Eoyal Society for the 12th December.] 



