760 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 124. No. 4. December 2012 
by placing radio transmitters on the ground and on 
canopy-level observation platforms at known 
distances and azimuths. We measured azimuths 
to transmitters from telemetry stations with 
known Universal Transverse Mercator (UTM) 
coordinates, and compared true azimuths to 
measured azimuths to calculate mean azimuth 
error. Wc used homing whenever possible to 
obtain visual locations of radio-marked hawks and 
minimize telemetry error. 
Wc located birds 1-3 times weekly by homing 
and conducted systematic searches using omni¬ 
directional antennas mounted on trucks to locate 
radio-marked hawks that had ranged outside our 
search area (Samuel and Fuller 1994). We 
recorded UTM coordinates when a radio-marked 
hawk was visually located with a portable Global 
Positioning System (GPS), recorded compass 
bearing to the hawk, and measured distance in 
meters with a range finder. Homing minimized 
location and azimuth errors from telemetry tests 
(~2 ) and allowed behavioral observations. We 
searched for radio-marked hawks during a random 
time period (0700-1100. 1101-1500, 1501-1900 
hrs) every other location to reduce location bias 
by time ol day. Wc generated UTM coordinates 
for all locations once fieldwork was completed 
(White and Garrott 1990. SAS 2001). 
Statistical Analyses.—We calculated the mini¬ 
mum number of locations for home ranges to reach 
an asymptote prior to analysis by randomly selecting 
ocations for six hawks with the greatest number of 
locations to create home ranges. We plotted each 
home range area by the number of locations to 
assess when home range area reached an asymptote 
We pooled years for home range estimates due to the 
relatively small sample size (n = 21) of radio- 
marked hawks available for analyses. 
All spatial analyses were conducted using 
rcView' 3.2 Animal Movements extension Vcr^ 
sion 2.04 (Hooge and Eichenlaub 1999). We 
estimated mean home range, mean weekly 
movements and core area changes for the 
breeding (15 Dec-30 Jim) and non-breeding 
(1 Jul-14 Dec) seasons. Wc also estimated 
aTm r T ,rave,cd from "-“PPing station 
or, r '7 7 1 brec ' di "« season cen,r °ids 
or all individuals during the study. Home ranges 
wfthl r a,ed US " li! " lc "sed-kemel estimator 
With least squares cross validation (Worton 1989 
Seaman and Powell 1990. Kernohan et al. 2001) 
T7 dt r m hreedin « centroids 
y paired Rcd-taiied Hawks were examined to 
learn if movements increased as energetic de¬ 
mands of nestlings and fledglings increased. We 
examined distances moved by juveniles to mea¬ 
sure dispersal distances following the post-fledg¬ 
ling dependency period. We used mixed-model 
A NOVA to test whether home range and weekly 
movements differed between age and gender 
(SAS Institute 2001). Seasonal core area shifts 
ol Red-tailed Hawks among breeding and non¬ 
breeding seasons were calculated using a multiple 
response permutation procedure tMRPP) in Pro¬ 
gram BLOSSOM (Cade and Richards 1999). 
We used Euclidean Distance analysis to assess 
Red-tailed Hawk patterns of habitat use (Conner 
et al. 2003. Bingham and Brennan 2004). We used 
the Animal Movements extension to generate 
1,000 random points within the home range of 
each hawk then overlaid each home range and 
random-point home range on the landcover of El 
unque and surrounding lands. We calculated 
distance from each hawk location and random 
point to the closest representative of each 
vegetation type by querying one vegetation type 
at a time (ESRI 2001). We used MANOVA to test 
didefences in means of the ratio vectors (Conner 
and Plowman 2001). Paired r-tests were used io 
test disproportionately-used habitats and a ranking 
matrix ol all possible pair-wise comparisons 
constructed to rank habitat types. All values are 
reported as mean ± SE (range); results were 
considered significant when P < 0.05. 
RESULTS 
We captured 32 Red-tailed Hawks in the 
Luquillo Mountains during our study of which 
27 were radiomarked and 21 were used for 
analyses (Table 1). A minimum of 25 locations 
was required tor home ranges to reach asymptote. 
Wc excluded two outlier individuals from analy¬ 
ses; a juvenile male (RTHA 13) and an unpaired 
second-year female (RTHA 7). The outlier 
juvenile male exhibited the greatest home range 
(28,791 ha), ranging beyond the Luquillo Moun¬ 
tains to the Sierra de Cayey, 40 km to the 
southwest (Table I). Red-tailed Hawks were 
monitored for an average of 5.9 ± 0.9 months 
(2-14), yielding an average of 46.7 ± 5.7 (24- 
110) locations per individual. 
We located hawks perched 62.3% ( 7 / = 628) of 
(he time, most frequently near the top of canopy 
emergent trees. Maximum distance from trapping 
■sites averaged 10,660 ± 1,291.3 in (range = 
6,875-17,490 m). Home range of Red-tailed 
