446 DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 
Mr. Kuichling. logical influence of forests, but the conclusions reached are often dis- 
cordant, owing to the obscurity of our knowledge of the various 
factors that enter into the problem. Some writers have laid stress 
on the mysterious properties of trees, singly and in groups, as exciters 
or conductors of electricity; others have given attention to the com- 
plex thermal influences of vegetable life; another class has studied 
the purely chemical actions of foliage; while a fourth class considers 
only the geological and hydrological relations, or the rainfall and run- 
off as commonly expressed and understood. These latter relations, 
however, involve all the others, so that in reality run-off is the resultant 
of a variety of physical forces, most of which are still almost 
unknown. 
The familiar factor called rainfall or precipitation is by no means 
easy to measure, or even estimate, with accuracy on a large drainage 
area. Its magnitude varies in almost every storm, both with the 
position and altitude of the rain-gauge and with the direction of the 
wind. Gauges placed only a few miles apart at the same altitude 
usually exhibit very considerable differences in the quantity of water 
caught, and these differences increase greatly in a mountainous region, 
especially on opposite sides of a high range. Still greater do such 
differences become when dealing with snow. In conjunction with 
one of the most skilful observers in the U. S. Weather Bureau, the 
writer once undertook to measure the depth of snowfall during two 
winters, but in no instance did the observations agree, either in respect 
to depth or weight, at several different points in the same city; neither 
did the differences balance even approximately in the course of several 
consecutive months. 
In view of these difficulties of estimating the approximate depth 
of atmospheric precipitation on a relatively small and nearly level 
-urban area with a number of favorably located gauges, it is easy to 
understand how much more difficult the task becomes in the case of 
large drainage basins presenting a great variety of surface and altitude, 
and having thereon only a few places of observation. 
Numerous cases might be cited in our oldest and most densely 
populated States, where important river basins with areas of several 
thousand square miles have only two or three rain-gauges, whose 
records embrace less than twenty years; and plenty: of cases occur 
where the observed run-off from adjacent tributary water-sheds of 
similar character in the same basin is widely different during a season 
of several months from apparently equal amounts of precipitation. 
Other instances also occur where the aggregate observed run-off 
during twelve consecutive months, from a mountainous drainage area 
of several hundred square miles, is greater than the reported precipita- 
tion thereon, measured at one or two stations. 
But while there are probably serious errors in the estimates of our 
rainfall, as well as a woful deficiency in the number of places of obser- 
