4 
islands. The great variability in amounts of rainfall may be conditioned 
by the shifting of cold-water and warm-water surfaces.” Knoch(l 927 ) an( ^ 
Wiens( 1962 ) state that while the strikingly low annual average precipita 
tion is surely connected with the temperature of the sea water, they 
do not consider that explanation satisfactory for the variations in 
precipitation, and relate this to wind direction and constancy instead. 
Howland is dominated by the south east trade winds with easterly 
winds prevailing most of the year. During the*Vinter" 
mwm 
period 
from October to March the winds are northeasterly and during the "summer 
n 
period (from April to September) the winds tend to be more southeasterly. 
While the northeasterly winds have been 
credited with the wet season 
Ifflffi on Howland (Brooks, 1926 ), latter data proves that the rains 
are more likely to be caused by the occuranee of W and NW winds. The 
confusion here could have resulted from the fact that the infrequent 
westerlies do seem to occur for the most part in the winter or northeasterly 
season., for in all recorded instances of shipwrecks on Howland (caused by 
weather) obtainable, squalls with westerly/winds were cited. And it would 
seem from this data that westerlies are most likely to occur during the 
winter or northeasterly season since 7 of the 9 shipwrecks occured during 
that period. Shipwrecks caused by westerlies on nearby Baker Island also 
occured mainly during season. 
Proof in the opposite direction for Howland’s rainfall resulting from 
westerlies is furnished by the low precipitation years 1910 , 1916 , and 
1917* During those years no disturbances of the trades by westerlies 
was recorded. (Enoch, 1927) 
