270 
PACIFIC SCIENCE, Vol. VIII, July, 1954 
given point. In one day’s time, however, it 
is possible for enough rain to fall to yield an 
above-normal total for the entire month. The 
high pressure, associated with this one day’s 
rain, would not compensate for the many 
days of low pressure associated with the 
drought, and the strong daily relationship 
would break down on a monthly mean basis. 
In other words, the problem stems from the 
fact that there are -no "'negative amounts” of 
rainfall to compensate for the few cases of 
exceptionally heavy rainfall. 
Thus, on a mean monthly map the point 
of maximum correlation may be shifted from 
the place where it appears on a daily map. 
The new position may be one at which high 
pressure favors persistent or repeated periods 
of rainfall but not, necessarily, concurrent 
rainfall. 
HILO (summer) 
Figure 5^ shows the correlation field of 
August 700-millibar heights versus Hilo rain- 
fall. The highest correlation on this chart is 
no more than one would expect to find by 
chance alone, and it is evident that Hilo sum- 
mer rainfall has little, if any, linear depend- 
ency on the mean monthly 700~millibar chart. 
This is probably due to the fact that trade 
winds over Hilo in summer are so deep and 
steady that their strength cannot be a critical 
factor and the fluctuations in rainfall must 
be associated with something other than the 
fluctuations in trade-wind strength. 
This very weak summer relationship is 
characteristic of all the rainfall indices tested, 
and thus far no alternative approach to the 
summer problem has led to very satisfactory 
long-period objective estimates. These low 
correlations indicate that Solot’s method is 
likewise inadequate in the summer months. 
HAMAKUA 
Fig. 5. Correlation field patterns of mean monthly 
700-millibar height versus monthly rainfall amounts. 
a, Hilo area, August, 17 cases; b, Hamakua Coast, 
winter months, 102 cases; c, Naalehu area, winter 
months, 102 cases. 
In contrast to the strong relationship shown 
by the Hilo winter correlation field (Fig. 1) 
is the much weaker relationship for the Hama- 
kua index (Fig. 5^). This rainfall index is 
