Traps number 29 and 30, in the shade of a large honey locust tree, also caught but few voles. Trap number 30 was only six feet from the base of the tree and caught but one vole throughout the study period. These two traps caught more _Peromyscus leucopus_ than any other pair, however, and both of them also caught pine voles (_M. pinetorum_). The area shaded by this tree permitted an extension of parts of the forest edge fauna into the grassland. In spite of the marked general tendency to avoid woody plants, some voles made their runways around the roots of blackberry bushes, sumac and wild plum trees. Some nests were found under larger roots, as if placed there for protection. More vegetation was found under the woody plants which the voles chose to use for shelter than under those which they avoided. It seemed probable that the actual condition avoided by voles was the bareness of the ground (a result of the shade cast by the woody plants) rather than the woody plants themselves. Running diagonally across the eastern half of the trapping plot in House Field there was a terracelike ridge of soil. On each side of this ridge there was a slight depression. Observations of the study plot in the growing season showed this strip to produce the most luxuriant vegetation of any part of the plot. Clip-quadrat studies confirmed this observation and showed the bluegrass understory to be especially heavy. This strip included the areas trapped by traps numbered 4, 5, 17, 18, 22, 23 and 37. With the exception of trap number 18, discussed above, these traps consistently made more captures than traps set in other parts of the plot. In winter, these traps also caught more harvest mice (_Reithrodontomys megalotis_) than any other comparable group of traps. Although the amount of growing tissue of plants probably is at least as important to voles as the total amount of vegetation, some correlation seemed to exist between the density of grassy vegetation and the density of populations of voles. A mixed stand of grasses, with an obvious weedy component, can support a larger population of voles than can either a nearly pure stand of grass or the typical early seral stages dominated by weeds. Probably the more or less continual supply of young plants provided preferred food easily available to voles. A more homogeneous vegetation would tend to pass through the young and tender stage as a unit, thus causing a feast to be followed by a relative famine. POPULATION STRUCTURE