1987 SCHNOES AND HUMPHREY: TERRESTRIAL COMMUNITIES IN FLORIDA 85

resulted from several factors. When pits were first abandoned, broad
areas were available for marsh development because water levels were
kept low by pumping. Over time water rose on the steep sides of the
pits, narrowing the zone for littoral vegetation. When shrubs and trees
grew on the bank, shading of cattails and other interactions may have
become important. Whatever the mechanism, older sites characteristi-
cally had very little emergent marsh vegetation.
 Lowest abundance occurred in the >30 year old class without lakes.
However, the difference between the unreclaimed areas with lakes and
those without was more evident in the species composition than in the
diversity and abundance measures shown in Figure 18. The strictly ter-
restrial sites were heavily dominated by P. gossypinus, and other species
were uncommon. The Christina site contributed much to the diversity
of the treatment because both species of shrews (Blarina carolinensis
and Cryptotis parva) and Mus occurred there. But these animals may
have immigrated from adjacent habitats-a residential community and a
grassy field. Woodrats were absent from these sites, except for one im-
mature individual caught at Christina during the summer quarter. This
animal also may have dispersed from neighboring locations.
 In contrast, the mature sites with lakes had considerably fewer Pero-
myscus, while moderate populations of Neotoma and Sigmodon were found
on most areas. This caused the difference in biomass between the two
treatments, because these species are much larger than the Peromyscus
that dominated the without-lakes community. Sigmodon probably was
favored by the retention of herbaceous cover (Fig. 12) along recreational
roads. Neotoma may have benefitted from the combination of mature
forest and the structurally diverse vegetation along water edges.
 Reclaimed Pastures.--In general the reclaimed pasture sites showed
low to intermediate small mammal community measures compared with
the other treatments (Fig. 18). These summary values are probably
somewhat higher than the "typical" case for reclaimed sites in the region,
because two sites, Kibler and H-4, accounted for 87% of the individuals
and 97% of the biomass on both reclaimed treatments (Tables 3-6). This
high concentration resulted from the dense populations of Sigmodon found
in the tall stands of grass that escaped mowing or intensive grazing. Most
of the modern reclamation treatments in the area more closely resemble
the other sites, but the Kibler and H-4 sites may indicate the potential
for succession if mowing and intensive grazing were to be reduced.
 The difference between grazed and ungrazed reclaimed sites was af-
fected by these two sites. The extraordinarily high numbers of Sigmodon
at the Kibler site resulted in very high abundance values, whereas H-4
consistently had four species present in more even proportions, resulting
in higher diversity for the grazed category.