988 
cases the amount of sunshine received at special times 
in the growth cycle of the plant will determine the 
commercial value of the fruits (an example is the 
August-September influence of sunshine on the sugar 
content of grapes). The rapidity of development de- 
‘pends on cumulative temperatures above a limiting 
value, that is, a certain degree-day value. The oc- 
currence during the growing period of a temperature 
below a critical value, such as the freezing point, can 
determine the partial or complete loss of a crop. Further, 
the proper water balance is of particular importance in 
plant development. There are times when too much 
water is as detrimental as is msufficient moisture at 
other times. The requirements within the life cycle of 
maximal and minimal amounts of water are very dif- 
ferent from one plant species to the other. The water 
balance is not determined by a simple climatic element, 
such as rainfall, but is governed by precipitation, trans- 
piration of the plant, and evaporation. This last element 
itself is a function of humidity, temperature, and wind 
speed. Thornthwaite has tried to establish some of 
these relations of evapotranspiration [83]. 
The literature abounds with studies of crop yields in 
relation to weather conditions [90e, pp. 293-476]. These 
relations can be used to establish facts about optimal 
growing conditions. In turn, these can serve to establish 
from long climatic records the risks to a given crop at 
various localities. A great deal depends here again on 
microclimatic conditions. This is particularly true of 
frost risks [47]. An area may, for example, be in general 
quite favorable for apple or peach orchards and yet 
some topographically poor exposures may be veritable 
frost holes and entirely unsuited for this type of culti- 
vation. 
Climatological principles have been applied to the 
introduction of plant seeds from one territory to another 
[66]. These studies of comparative climato-ecology are 
still in their infancy and amenable to much improve- 
ment. The difficulties involved have already been 
pointed out (see page 980) and allusion was made to the 
problem of climatic imfluences upon pests. 
Experience has shown that a definitive fire weather 
exists, especially in forests. For effective fire-fighting 
procedures it is useful to estimate the danger class of 
each day and season. This will result in better dispositions 
for lookout posts, fire-fighting equipment, and per- 
sonnel. The climatic information serves both for plan- 
ning purposes and for the anticipation of fire-weather 
stations. These are later supplemented by special 
weather forecasts. 
For many forest areas of the United States special 
jire-danger meters have been worked out. These are 
composite slide rules which have been described in the 
literature [41]. A case history for an urban area may, 
however, be of interest. It shows the technique of ar- 
riving at a fire-danger scale. This scale was derived for 
grass and rubbish fires in the city of Raleigh, N. C. 
It was assumed that the immediate causes for such 
fires, namely flying embers, carelessly dropped burning 
matches and cigarettes, etc., remain essentially con- 
CLIMATOLOGY 
stant so that weather conditions contribute most to the 
variability of incidence. 
The climatic elements mainly entering into a classi- 
fication of fire hazards are precipitation, relative hu- 
midity, and wind speed. Precipitation and relative 
humidity are important for the period preceding the 
fire, while the wind speed enters at the time of ignition. 
For establishing the scale, the conditioning and co- 
existing weather conditions were noted for 90 days on 
which three or more grass or rubbish fires were observed. 
It is not surprising that precipitation and humidity 
values immediately preceding the fires were more im- 
portant than those further back in time. This required 
weighting factors for the sequences of these elements. 
These were simply obtained by multiplying the values 
on the jire day (three or more rubbish or grass fires) 
by 4, on the preceeding day by 3, on the second pre- 
ceeding day by 2, and by using the actually-observed 
values for the third day prior. 
The statistics showed immediately that 70 per cent 
of the fire days had a relative-humidity factor of less 
than 40 per cent, while only 18 per cent of all days had 
a humidity factor of less than 40 per cent. Over 80 
per cent of the fire days also had a weighted precipi- 
tation factor of 0.01 m. rainfall or less (of all days only 
37 per cent have a precipitation factor of less than 
0.01 in.). The simultaneous wind speed did not show 
quite as good a discrimination, although 88 per cent of 
the days with five or more fires had wind speeds of 18 
mph or more. However, 75 per cent of all days have 
winds exceeding 12 mph at one hour or another. 
These statistics resulted in the following scale: 
Weighted Weighted - 
relative precipi Wind 
humidity tation speed 
% in. mph 
Good fire days.........=40....and....=0.01 and.... 513 
Poor fire days......... 560....and....50.05 
eerie or570 
oes (Oras tae eed alec coe SS Ocan) 
Days that did not fall into the classes good or poor 
were classified indifferent, that is, the weather would not 
contribute to the fire hazard nor would it definitely 
retard fires. The discrimination of the scale is shown for 
four test months (January—April = 120 days) for four 
years (total 480 days) (Table VII). 
Tasie VII. Frre OccURRENCES AND FIRE DANGER CLASS 
Fire Class Total days | Total fires race per 
Goodthretdaye nse eee 110 226 2.05 
Indifferent fire day............. 224 193 0.86 
12 OE INE CANscesccescuagons0ase 146 32 0.22 
The table shows that the scale is quite satisfactory. 
The good fire days had over nine times as many fires as 
the poor fire days. 
This group of problems is also peculiarly the province 
of the synoptic-climatic technique of approach. An 
example is the establishment of a time schedule for 
