1987 SCHNOES AND HUMPHREY: TERRESTRIAL COMMUNITIES IN FLORIDA 61


measured along the transect with a line-intercept method. Herbaceous vegetation was
sampled using ten 20 cm x 50 cm plots, located at 10 m intervals along the transect. Plants
were identified to species level, and percent cover in each plot was visually estimated.
Plant names agree with those used by Radford et al. (1964) and Long and Lakela (1976).
 Tree species were sampled using 20 of 40 cells, each 10 m long and 5 m wide, by
selecting every other cell on each side of the 100 m transect. In each cell, trees (>5 cm
dbh) were identified and counted. Height was measured with an optical clinometer, and
diameter at breast height (dbh) was measured with a tape measure. Basal area (BA) was
computed for each species using the formula

 BA = D
 4

where D = dbh. Frequency of occurrence for each species was obtained for the 20 cells.
Relative frequency (RF), relative abundance (RA), and relative dominance (RD) for each
species were calculated using the frequency of occurrence, number of stems, and basal
area respectively compared with the totals for each transect. An importance value (I) was
then assigned to each species such that I = RF + RD + RA.
 SMALL MAMMALS.-Small mammal communities were sampled once each quarter using
both Sherman live traps and museum special snap traps. A removal technique was em-
ployed and all animals caught were identified, weighed, and measured, and age class and
reproductive status were noted. Each site was trapped for three consecutive nights using
64 stations consisting of one live trap and one snap trap baited with peanut butter and
oatmeal. These stations were arrayed in two 4 x 8 grids, with trap stations 8 m apart. Traps
were checked each morning and rebaited as necessary.
 BIRDs.-Avian communities were sampled using a variable-width transect method
 (Emlen 1971). Transects were walked through study sites during the first 2.5 hours after
 sunrise, and all birds seen or heard were tallied, along with the distance from the transect
 when first detected. Birds were excluded if they occurred beyond 100 m of the transect
 line. Transect counts were standardized by extrapolating to the equivalent of a 1 km tran-
 sect. In all but the first quarter, time permitted transects to be run twice on different days
 on each site. During autumn quarter, only one run was conducted.
 Following Emlen's method, a coefficient of detectability (CD) was calculated for each
 species, each type of habitat, and each quarter. These coefficients were then used to
 correct the original count data to provide a more reliable estimate of the number of birds
 present within the transect boundaries. A separate set of CDs was computed for each of
 the 8 treatment classes, except that the two types of reclaimed pastures were considered
 as one group.
 Biomass estimates were made from several sources, including the University of Flor-
 ida bird collection, Clench and Leberman (1978), Greenewalt (1975), and incidental weight
 measurements made on birds during field work.
 INSECTS.-Insects were sampled using light traps with 8 watt blacklight bulbs (BioQuip
 Inc., 1320 E. Franklin Ave., El Segundo CA 90245). Light traps were set 1-2 m above
 the ground in the evening and retrieved early the next morning. The samples were oven-
 dried for 48 hours at 700C and then were weighed to yield biomass data. Time permitted
 only 1 set of samples (those from winter quarter) to be sorted to Order. The diversity
 values used in this report are from that quarter.
 Only three light traps were available for use, and technical problems at times reduced
 that number to two or even one. This caused the sampling of all sites to be spread out
 over at least ten days and at times much longer. This factor added to the high variance in
 the data expected from the nature of insect hatches and insect responses to weather vari-
 ation.
 HERPETOFAUNA. -Original plans called for pitfall-with-drift-fence trap arrays to be