394. QUEENSLAND AGRICULTURAL JOURNAL. [1 May, 1902. 
(d) Part soaks into the soil up to saturation point, 
(¢) Part is absorbed by plants for their growth and transpiration, 
(f) The remainder sinks to lower levels, where it either forms subter- 
ranean reservoirs or percolates till it again comes to the surface as 
springs. 
Calling R the total rainfall— 
R= at+b+c+dtetf. 
Ona level plaingthere is no surface flow, and the equation becomes— 
f= R-(a+b+d+e). 
+ _ In the simplest case, that of a perfectly bare plain, a, c, e become zero, 
and the equation is— 
f= R-—(6+4). 
Of all these factors the only one that can be measured with anything like 
accuracy is the portion a evaporated off the trees. Even this factor, according to 
the Mariabrunn observations, is liable to very considerable errors of deter- 
mination, since no two rain-gauges will give the same readings even under the 
same tree. It is necessary to employ a large number of rain-gauges, including 
some embracing the trunks of the trees. Even employing 20 rain-gauges, 
totalling 10 square feet of opening, the probable error is at least 1 per cent. of 
the fall. The measurements should be made either (1) for each individual 
shower, (2) for a long series, (3) by grouping the showers according to their 
intensities. 
It has been found in Europe that a broad-leaved forest prevents 1 to 8 
tenths, and a conifer forest as much as 5-tenths, of the total precipitation from 
reaching the ground at all. But these figures hold aa only for those 
localities where they were obtained. In countries where the rainfall is heavy 
and continuous, the forest soil in any case will be about as thoroughly watered 
as a bare soil. 
All the other fractions, b, c, d, e, f, composing the total fall, are still ve 
undetermined in forests and other lands alike, and they vary so much wit 
every possible local difference that they are hardly likely ever to be capable of 
satisfactory measurement. ; 
fe: Reasoning from the known facts that, in spite of obstruction by the 
crowns, the forest soil is as well watered as the soil outside, and that the 
evaporation in a forest is much less as proved by the greater moisture of the 
surface soil, it was supposed that forests contributed more than anything else 
to the maintenance of subterranean supplies (level plains are still referred 
to). The results obtained in Russia, therefore, came as a great surprise. 
Soundings taken during the growing season’ (Ist June to Ist Septem- 
ber) inside and outside the forest of Chipoff (Government of Worone}j) 
showed that the water level below the forest was some 32 feet lower 
than outside. In the Black Forest (Government of Kherson) the level 
was some 12 to 16 feet lower. Presumably these figures are extremes for the 
following reasons:—(1) The measurements were taken at the season when trans- 
piration is greatest; (2) they were made in localities where the rainfall was only 
12 inches in the year, where there was as a probable natural consequence of the 
dryness an almost complete lack of natural forest, and where consequently the 
forest would have to pump all it could in order to maintain itself. An increase 
of forest area would probably reduce the necessity for so much pumping by the 
roots. ‘The experiment was repeated by M. Ototzky much farther north, under 
the 59th degree of latitude, in the Government of St. Petersburg, where the 
climate is cooler and moister and the rainfall averages 20 to 30 inches. ‘The 
subterranean water is plentiful, yet again the forest lowered the level, but this 
time only by 20 to 46 inches. In order to check the Russian results, an 
experiment has been started by the French forest officers in the forest of 
Mondon, near Luneville. It is a level forest of about 5,000 acres, situate on a 
