RESEARCH METHODS IX STUDY OF FOREST ENVIRONMENT. 29 
Granting that a full and satisfactory record of soil temperatures 
may be obtained by the use of the thermograph, it may still, because 
of the cost of this instrument, be impossible to obtain the desired 
comparison of a number of sites. The best alternative would seem 
to be to make one thermograph serve for a number of stations by 
placing it successively at the several stations until the nature of the 
diurnal oscillation, for a given season, has been worked out for each 
station. These oscillations will depend so greatly on the character 
of the insolation, that a curve for one point could hardly be expected 
to apply at any other point. With a mean daily curve, however, a 
single thermometer reading each day may give a very good basis 
for approximating the mean soil temperature for the day. If this 
is convenient, the reading may be timed to accord with the most 
probable hour for the mean temperature to occur. 
With hourly soil temperatures for a period of a week at any 
season, tabulated on the " Hourly (Air, Soil, or Actinograph) Tem- 
peratures " form, the mean hourly temperatures may be computed, 
as well as the mean for all of the days concerned. From the former 
may be obtained a correction factor for any hour, which, added to 
the reading for a similar observation hour will give approximately 
the mean temperature. 
For instance, a study of station A-l at "Wagon Wheel Gap, Colo, 
(steep northerly exposure), at midsummer, showed that the daily 
oscillation was about 1.35°, that the mean temperature Avas ap- 
proached very closely at 7 a. m. or 7 p. m.. the minimum not occur- 
ring until 2 p. m.. and that the correction for a 9 a. m. reading, on 
6 days, varied from -j-0-iO° to +0.50°. with a mean correction of 
0.34. Similarly at station A-2 (south exposure), it was found, that 
the aproximate mean would be read at 5 a. m. or 4 p. in., that the 
minimum occurred at noon, that the daily oscillation was 2.37°, and 
that a a. m. reading must be corrected by —0.88° to give the mean 
for the day. Corrections for six individual days varied from —0.50° 
to —1.35°. 
Moore (11) states that at a deptli of 3 feet daily oscillations are 
not felt. It is believed that they are, as a rule, too small even at 
2 feet to warrant consideration, although in excessively insolated 
soils the procedure described for 1-foot temperatures may be fol- 
lowed. 
Another method which suggests itself for determining the probable 
variation from the mean of any daily temperature reading at a fixed 
hour is to compare the annual mean temperature at the shallow 
depth with the mean for 4 feet or greater depth where, it may be 
assumed, the daily values are not affected by regular oscillations. 
For, while at any time the deeper soil may be cooler or warmer than 
the surface, the deeper soil always evincing a definite " lag " when 
