88 
PACIFIC SCIENCE, Vol. Ill, January, 1949 
lO 
o 
to 
o 
in 
O 
to 
CXI 
OJ 
ro 
ro 
si" 
CD 
CD 
CD 
CD 
CD 
Fig. 2. Graph showing changing proportions of east wind (bottom) and northeast wind (top), by 
the year, from 1905 to 1947. The residue (blank) of less than 20 per cent includes all other wind 
directions. 
This plotting gives a clear indication of swing 
but some angular scale is desirable. This has 
been developed by regarding the proportions 
of wind in the two major sectors (northeast and 
east) and in the two adjacent ones (north and 
southeast) as representing approximately a nor- 
mal curve of frequency distribution. By very 
approximate account of the changes in the lesser 
proportions with the change in the major, it 
has been developed that 10 units of change in 
the difference between northeast and east per- 
centages is roughly equivalent to 5 degrees 
change in direction of the prevailing wind or 
the central position. For the present purpose, 
and to give some sense of direction to the graph 
and the arrows as drawn, this relationship is 
sufficient, though rigorous calculations or actual 
measurements through a period of some years 
would possibly indicate a moderate error in this 
ratio of comparison and also might show that 
the ratio is not quite uniform throughout. 
The data represented in Figure 1 leave little 
doubt that a systematic shift is in progress and 
that it is cyclical in character. At the eastern 
extreme, the successive positions of arrows also 
suggest the swinging of a lesser cycle on the 
major one, a rather common state of affairs. 
Since the major cycle is not complete, any esti- 
mate that might be made as to period must be 
very rough. Perhaps we can go no further than 
to note that the three quarter-periods between 
the approximate points 1908-1920-1934-1942 
are 12, 14, and 8 years, suggesting a whole 
period of the order of 45 years. No basis is 
known for such a period and the writer desires 
only to state roughly the condition that is indi- 
cated by the data in Figure 1. 
The purpose of this note is to show that there 
is a definite, rather smooth, cyclical swing which 
will apparently complete one whole period from 
1908 within the 1950 to I960 decade. To de- 
termine its cause will require more knowledge 
of other cycles than the writer, or perhaps any- 1 
one else, now possesses. To determine its true 
period will of course require a much longer 
period of record, and the nature of the various 
wind patterns that enter to produce the fre- 
quency distribution could be better analyzed if 
a more detailed record could be obtained over 
a number of years. 
