490 BOTANICAL GAZETTE [JUNE 
11 and 13, the latter of which was placed near the border of the 
association, while the former was in the center. With the estab- 
lishment of the mixed forest association the sand area reaches the 
present culmination of the successional series. Throughout this 
series the amount of evaporation has steadily decreased, reaching 
its lowest rate in the association genetically highest. 
Actual observation shows that the succession of the bunch-grass 
by the Quercus velutina association begins in the normal association, 
where the rate of evaporation is high. It also shows that members 
of the mixed forest association appear in the Quercus velutina asso- 
ciation while the evaporation rate there is still relatively high. In 
both cases the development of the association reduces the evapora- 
tion. Succession, therefore, does not depend upon evaporation; it 
is a cause rather than an effect. 
The other records shown in the diagram are not to be placed in 
this genetic series. Sufficient instruments were not at hand, nor 
does this particular locality give the opportunity for investigation 
in the series of associations beginning with the open water and 
ending with bottom-land forests. Nevertheless, one or two points 
are brought out, notably that the evaporation on the open sand of 
the river beach is not nearly so great as that from the open sand 
of a blowout. The river is here a most important factor, tending 
to reduce the extremes of temperature during the day, and main- 
taining a relatively high humidity during the night. Evaporation 
from the willow and maple islands was greater than from the mixed 
forest, although the latter grew in dry upland sand, and the former 
in mud, during part of the experiment covered and at all times 
surrounded by large bodies of water. This is simply another 
instance indicating that the vegetation is of greater importance 
than the physical environment in controlling evaporation. 
Conclusions and summary 
t. Since proper conditions for maintaining an absolute standard 
were lacking, the results were all calculated on the basis of one 
instrument arbitrarily chosen as a standard, and are represented 
in relative terms only. 
