WASSMER ET AL.: SOUTH-CENTRAL FLORIDA BOBCAT ECOLOGY


 The validity of this method of density estimation rests on the assumptions that (1) male
home ranges do not overlap appreciably, (2) male home ranges are essentially contiguous, with
little or no unoccupied areas between them, and (3) female ranges are included within male
ranges. Although these assumptions were probably not fully met, the available data suggest they
generally applied on the core area. Any bias in the method is probably in the direction of an
overestimate of density, particularly for the entire study area. On the other hand, using the
alternative method of dividing the total study area (as determined by maximum movements of
individuals captured and marked on the core area) by the number of bobcats believed to be
present would probably significantly underestimate density because of the increasing probability
of not detecting the presence of bobcats at increasing distances from the core area.
 Locations of marked individuals obtained by radiotracking, capture, or visual observations
were plotted on a large scale (1:4800) map of the study area gridded at 200 m intervals into 4-ha
quadrats. The 4-ha quadrat size was selected as the minimum area of resolution based on the
accuracy of radio fixes at distances of 1 km or less. Home range areas were estimated by the
minimum convex polygon method (Mohr 1947). For purposes of this study, the area used by an
individual based on all locations over the entire period it was monitored is termed the "overall"
home range, and areas used by individuals during one of the 12 time periods mentioned above
are designated as "short-term" home ranges. The distribution of fixes among the 4-ha quadrats
enclosed by home range boundaries was used as a measure of the intensity of utilization of home
range areas (Siniff and Tester 1965).
 Estimates of daily movements were based on straight-line distances between the first radio
locations on consecutive days of monitoring.
 Whether a bobcat was moving ("active") or stationary ("inactive") was determined by the
 nature of the signal and the animal's location on successive fixes. When more than one fix was
 available for a bobcat in a given hour, the first record was arbitrarily selected to indicate the
 animal's activity state during that hour. The period between 2400 and 0600 hours was relatively
 poorly sampled, with only 325 (7%) of the total of 4966 hourly records included within this
 interval.
 The amounts of different habitats contained within home ranges of individual bobcats were
 determined by counting the number of 4-ha quadrats containing each type. When more than one
 habitat occurred within a quadrat, each was assigned a fractional value corresponding to the
 number of habitat types represented, regardless of the actual coverage of the habitats in the
 quadrat. Habitat preferences were assessed by comparing the observed frequencies of locations
 in quadrats representing different habitats with expected frequencies based on the relative areas
 of those habitats within the home range. Fixes within quadrats containing more than one habitat
 were arbitrarily assigned fractional values for use of each habitat corresponding to the number of
 habitats represented within the quadrat.
 Statistical tests used include Mann-Whitney U (U), Spearman rank correlation (Rho), and
 Chi-square (X ) as described in Siegel (1956) and the Log-likelihood ratio (G) given in Zar
 (1984).



 RESULTS AND DISCUSSION

 Capture Success



 Overall capture success (18 original captures, 19 recaptures) was 1.8
 captures/100 TN. The number of TN in different seasons ranged from 609 to
 660, and capture success (per 100 TN) was as follows: December through
 February, 3.1; March through May, 1.1; June through August, 0.7; September
 through November, 1.1.