12 BULLETIN 1059, U. S. DEPARTMENT OF AGRICULTURE. 
fined. The use of these same data in strictly local studies will de- 
pend entirely on the minute examination of the conditions surround- 
ing the station. 
KNOWLEDGE OF EXISTING WEATHER BUREAU STATIONS NECESSARY. 
Before attempting any rueterologieal observations the investigator 
should visit the nearest permanent meteorological stations and ob- 
tain a clear understanding of the manner in which the observations 
are made, compare his own instruments with those of the station, 
and ascertain the natural conditions in which the permanent station 
is located and the extent to which they are typical of the region. 
This is essential to enable the investigator to decide whether and to 
what extent he would be justified in connecting his special meteoro- 
logical observations with those of the permanent station. Observa- 
tions at permanent, well-equipped Weather Bureau stations are not 
always conducted in the way that meets the special needs of the 
investigator. There may be observations essential to the forester 
which are not being made at all. Furthermore, the data of the per- 
manent station will not always enable one to judge of the effect of 
the climatic conditions upon forest vegetation. For instance, .the 
measurements of the temperature of the air are always made at a 
regular TVeather Bureau station at some height above the ground 
and in a more or less open place outside of the forest : while to the 
forester, the temperature of that layer of the air in which most of 
the forest vegetation is found has the greatest significance. Again, 
while a very precise measure of precipitation may be of no use to 
the investigator, the amount falling in single storms may vary so 
greatly in short distances that a record obtained a few miles away 
will be very misleading. It is thus evident that forest research has 
special meteorological problems, and that usually the long-estab- 
lished weather station may ^erve better as a control than as a definite 
point for obtaining information about forest conditions. 
COMPUTATION OF ALL WEATHER DATA BY PERIODS OF GROWTH 
AND REST. 
One essential thing to be kept in mind is that plants may react 
to the climatic condition- in altogether different ways during periods 
of growth and rest. To analyze the reactions of plant life it is 
usually desirable, therefore, to :ompute climatic data by such 
periods. They may be based either on a knowledge of the particu- 
lar plant formation which each observation-point represents, or on 
the average period of the native vegetation of the locality. I sually 
it will be preferable to adopt first a " growing seasoB " for the whole 
region under study. Later, for more exact comparison of the com- 
ponent formation- and after careful determination, the specific 
