2 
average may fall in one month. Because of this variability the general 
aspect of the island is much drier than the average annual rainfall would 
suggest. The growth of the vegetation may vary so much that descriptions 
written several years apart may scarcely appear to refer to the same island. 
In commenting on rainfall in the south Pacific, Seelye (1950) wrote 
that annual rainfall is most variable along the western tongue of the equa¬ 
torial dry zone which has a steep north-south rainfall gradient. Thus com¬ 
paratively small disturbances of the controlling atmospheric circulation would 
be expected to produce spectacular changes in that area. Pacific Project 
observations on Howland, which is in this area, indicate that this conjecture 
is correct. When rain does fall, it usually falls in torrential showers, 
and several inches may fall in a few hours. July and October, 1963* were very 
wet but there seems to have been little rainfall since that time, judging 
from the state of the vegetation. 
Ramsay (1925) of the Whippoorwill Expedition doubts that Howland has 
a mean annual precipitation of over three inches, while a Pacific Missile 
Range publication states that Howland’s annual precipitation may be estimated 
at twenty-five inches. Knoch (1927) apparently was more accurate when he 
simply stated that the annual precipitation may vary as much as twenty-four 
inches from y©ar to year./ 
the dividing line between the two tradw winds is narrowest, but according 
