1905] BROWN—THE PLANT SOCIETIES AT YPSILANTI 279 
The trees are mainly nut-bearing and ripen their fruits, unfold 
their leaves, and bloom relatively late. Thus we commonly see a 
tamarack swamp or a floodplain forest with young foliage well along 
before the surrounding oaks have begun to show marked signs of life. 
Both curves show that the highest per cent. of vernal as well as north- 
ern forms occurs in the floodplain societies and on the juniper slopes. 
The highest per cent. of southern forms occurs along the favorable 
slopes of the bluff and in the aquatic societies. 
CONTROLLING FACTORS. 
1. CLimatic.—A number of species have been pointed out whose 
distribution is more or less restricted to the stream, although this 
will be shown to be less true near the mouth. Most common of these 
are the hickories, walnut, yellow oak (Q. acuminata), red elm, silver 
maple, red and green oak, sycamore, hoptree, Rhus aromatica, Juni- 
perus communis, J. virginiana, Iris versicolor, Meibomia canadensis, 
M. grandiflora, Ambrosia trifida, Dioscorea villosa, Sanguisorba 
canadensis, and Mesadenia atriplicifolia. Most of these are near 
the northern limits of their range. Since many of these forms occur 
in a wider range of habitat in the south, it may be assumed that cli- 
mate plays some part independent of other factors in causing such 
Species to become limited to the neighborhood of the stream. 
2. ATMOSPHERIC.—The fact that the forest is usually open along 
the edge of the bluffs seems to account in large part for the presence 
here of the juniper and associated species. Their behavior indicates 
€xacting light requirements, as they thrive only in open situations, 
and early show a decline when shaded. To a less degree the sumac 
Society occupies open places in a similar manner. 
3. MECHANICAL.—The mechanical effects of stream erosion are 
continually bringing about changes of topography and soil to which 
the plant societies are in turn conforming. Although a number of 
luffs are being denuded by the cutting of the current, the floodplain 
areas are not likely to be very greatly increased; since it has been 
shown that a stream cuts within the limits of a belt not over eighteen 
mes the width of the stream at that point.’ The fact that Populus 
‘ 
— JEFFERSON, M. S. W., Limiting width of meander belts. Nat. Geog. Mag. 
‘373-384. 1902, : 
