3 
Ttve attempts at explanation concern themselves chiefly with the 
low annual amount overall, which does not seem to fit thsxtte into the 
tropical rain belt at all. Woeikoff (l88o) says on this point that the 
Howland area is the region where one assumes that the 
dividing line between the two trade winds is narrowest, but accord 
,• 3,. *6 * 
to actual observations by Hague in 1862, the two trade winds must have 
no dividing line here and merge completely MM into each other. Tha J 
would thus belong year round to the low precipitation zone, particularly 
lacking mountains or other local conditions that might cause rain. In 
fact, the very flatness is credited by some with the actual ability to 
discourage rain. According to Hague's 1862 observations "a shower of 
rain approaching the island divided into two parts and passed by to the 
north and south, the cloud being split by the heated column of air that 
rises from the white coral sand.". Ellis(1937 )> Bryan(l942), and Ramsay(l924) 
also mention this phoenomenon. This would tend to explain the fact stated 
by Bryan and Hague that it usually rains at night on Howland if at all. 
POBSP observations have also noted this fact. 
W. Koppen (1923.) approaches the matter of rainfall variation in this 
region from the oceanographic facts. He says, " As an explanation we can 
deduce here only the strip of cold water which (probably sucked up from 
the depths by the strong westerly stream in the southeast trades) runs 
from the Galapagos along the equator almost,(but not quite) tb these 
