142 BOTANICAL GAZETTE [FEBRUARY 
of 65% miles. This falls short of the real fact, because in Jordon and 
Cadiz, in the absence of certainty about some former creeks, we have 
omitted them; and because in exceedingly dry times the water has 
disappeared further than we represent, and these are the crucila 
times. : 
With the drying up of their head waters, the flow of the creeks has 
of course diminished everywhere. There are no measurements of flow 
at any time in the past, and, having nothing to compare with, we have. 
made none. The millponds show the state of affairs most graphically. 
Rich, Honey, and Skinner Creeks each ran several mills; altogether 
there were twelve of these. All have been discontinued or equipped 
with steam power. On each creek there were running by water power 
twenty years ago at least two mills. The occasion for saw and grist 
mills is largely gone—reason enough why some of them are absent; 
but several now use steam. One mill near the mouth of Skinner Creek 
uses both water and steam. Where most of these mills stood less water 
runs now than used to be wasted. 
A change in the forest covering of a region might indirectly affect 
the stream flow, through an effect on the total rainfall or its seasonal 
distribution; we have no evidence that this occurs. And the more 
direct effect in changing the structure and exposure of the ground 
might be years in making itself fully seen. Still, because of the length 
of the time involved, and because the deforestation had been relatively 
slow for some time preceding twenty years ago, we conclude that the 
creek flow there was adjusted to the forest area and condition; like- 
wise, there having been no considerable change within five years, the 
timber and creeks seem to be in equilibrium now. If this be so, and 
assuming still that these four towns fairly represent the general region 
near about them, we can conclude that 27 per cent. of forest will nearly 
enough conserve the surface flow of water in country originally 83 pe 
cent. forest so that the difference does not appear in any considerable 
loss of springs or drying up of creeks; but that if this forest be further 
reduced to 6 per cent., many creeks dry up and the flow of water in 
those remaining is reduced to probably about one-third. 
About the rainfall in Monroe nothing exact is known, and if we 
had data, their relation to the deforestation would be very doubtful. 
While the loss of flowing water is a very local function of deforestation, 
the rainfall, if it be a clear function, is not localized at all. And the 
‘Variation in different seasons and localities is so great that very exten” 
sive data in space and in years, as to both rainfall and the timbered 
