In November, 1949, the rats were almost gone from the area of north slope and hilltop in oak-hickory-elm woodland where the most intensive live-trapping and other field work had been done the previous year. The following descriptions of houses remaining on the area at that time give some idea of the habitat, and of the course of events correlated with the fluctuations in numbers of woodrats. No. 1. At the hilltop outcrop, partly on a substrate of limestone boulders, built around an elm of two-foot DBH, which lent support to one side. A hackberry sapling one inch in stem diameter grew through the middle of the house, providing further support. The house was two feet high and six feet in diameter, and was in obvious disrepair, with a hole several inches in diameter in its top. It had been occupied in the autumn of 1948. It was constructed mainly of sticks, ranging in diameter from approximately one inch to straw size. Many of the sticks, from .4 to .5 inches in diameter and one to two feet long, seemingly would have been heavy burdens for a rat, although they were of light-weight wood, sumac and elm. Mixed with the sticks were quantities of dry leaves, bark, and chips of wood, all material appearing old and weathered. This house was in elm-oak-hickory woods 50 feet from a cultivated field on the hilltop to the east and south. To the north and west the escarpment sloped away abruptly. There was a coralberry thicket beneath the trees on the adjacent hilltop. [Illustration] FIGURE 1 (A) Map of part of University of Kansas Natural History Reservation, showing first-capture sites for all woodrats live-trapped in the autumn of 1948. Because of the short time involved and the few traps available, much of the area shown was not thoroughly trapped. Woodrats were abundant, though much less so than in 1947, as shown by the large number of deserted houses. (B) Map of woodrat study area, same as shown in (A), showing first-capture sites for all woodrats live-trapped in 1949. Woodrats were still moderately abundant, but much below the level of the previous year. Triangles indicate those capture sites not sampled in 1948. (C) Map of woodrat study area, same as shown in (A), showing first-capture sites for all woodrats live-trapped in 1950. Numbers were medium-low, having undergone drastic reduction from the peak level. Triangles indicate those capture sites where trapping was not done in earlier years. (D) Map of woodrat study area, same as shown in (A), showing first-capture sites for all woodrats live-trapped in 1951. The population was low, but had not yet reached its lowest ebb. (E) Map of woodrat study area, same as shown in (A), showing first-capture sites for all woodrats live-trapped in 1952, when the population had declined to relatively low numbers and disappeared from much of its former habitat.