1987 SCHNOES AND HUMPHREY: TERRESTRIAL COMMUNITIES IN FLORIDA 63

occurred, though for most variables there was a strong seasonal effect. Therefore the
analyses of seasonal and seasonal data were identical. Distinctness of final land uses was
first tested by applying Duncan's multiple range test to treatment means. Natural con-
trasts anticipated in the experimental design were confirmed as groups by the Duncan's
test, so the more powerful analysis of variance specifying these contrasts was conducted
(combined over seasons for seasonal data) to look for more subtle distinctions between
contrasted treatments. Contrasts were as follows: mature unreclaimed sites with vs. with-
out lakes, ungrazed vs. grazed pastures, clay vs. mature unreclaimed sites, and clay vs.
reclaimed pastures. Significance of hypothesized age effects was established by pooling
sums of squares of non-significant responses in with the error sum of squares until only
significant effects remained. The effect of site age was modelled by computing equations
of significant variable responses to site age, using linear regression of the transformed
data. Finally, simple relationships between animal and plant variables were sought by
stepwise multiple regression.

 RESULTS

 VEGETATION

 CONSOLIDATED CLAY SETTLING PONDS.-The three sites were cho-
sen to exhibit a successional sequence within the treatment-the Swift
site supporting the youngest community (Fig. 1) and A-3 having the
oldest (Fig. 2). This sequence was reflected in the vegetation analysis
(Table 2). The Swift site was dominated by a dense, young stand of willow
(Salix caroliniana) and a few Baccharis, with a well-developed unders-
tory of grasses. This site was wet enough that cattails (Typha sp.) and
rushes (Juncus sp.) also were established in places. Most willows in this
situation were too small to be considered trees (dbh <5 cm) and thus
were measured as shrubs.
 The intermediate-age settling area, N-2, also was dominated by wil-
low, but these were much larger than in the Swift site. Willows of tree
size in N-2 had a mean dbh of 6.3 cm and a mean height of 4.4 m. Herb
cover, shrub cover, and shrub diversity were somewhat higher than at
the Swift site. However, willows still heavily dominated the shrub layer.
 Tree-size wax myrtle (Myrica cerifera) dominated A-3, the oldest set-
tling area, but willows were still well established. These willow trees
were generally larger than the wax myrtles and occurred in wet depres-
sions. In addition, numerous dead trees all were found to be willows.
Vines dominated the shrub and herbaceous cover, but this may be mis-
leading because the crowns of the abundant, young wax myrtles were
intermixed with the crowns of the mature trees, and it was impossible to
determine what percent of the cover was contributed by the younger
trees. In this case only shrubs less than 3 m in height were measured.
No herbaceous vegetation was present. This late successional community
on clay settling ponds is unlike undisturbed mature vegetation found in