Houston et al. • HOME RANGE SIZE OF BREEDING TURKEY VULTURES 
473 
METHODS 
Study Area— We (racked vultures at four sites 
in central Saskatchewan, Canada: two in the 
Aspen Parkland ecoregion (i.e.. a transition zone 
between southern dry grasslands and northern 
boreal forest) and two in Lite Southern Boreal 
Forest ecoregion (below and above 52.5 N 
latitude). Most Aspen Parkland is now cultivated: 
however, native grasslands and woodlands, re¬ 
spectively. are dominated hy fescue grassland 
( Festuca spp.). and quaking aspen (Populus 
tremuloides) and balsam poplar (P. balsamifera). 
The Southern Boreal Forest ecoregion is the most 
diverse biotic region in Saskatchewan (Smith 
19%. Thorpe 1999) with an overstory of mixed 
deciduous (aspen and balsam poplar) and conif¬ 
erous trees (white spruce [Picea g/auca], black 
spruce \P. mariana ], and jack pine [ Pinus bank- 
si ana]). 
Nest Location and Radio Transmitters. —We 
used radio and newspaper publicity to contact 
farmers who believed that Turkey Vultures nested 
on their property. This was critical as many of the 
farm buildings occupied by nesting vultures were 
"hidden” by aspen and caragana ( Caragana spp.) 
trees growing in the long-deserted farmsteads. 
We captured six adult vultures at their nests and 
equipped each with an alpha-numeric patagial tag 
on the left wing and a back-pack style (Steenhof 
et al. 2007), solar-powered Global Positioning 
System (GPS) satellite transmitter. We attached 
four units (#’s 57952, 57953, 65544, and 65545 
|Microwave Telemetry, Columbia, MD. USA]), 
which provided locations almost every hour, to 
one of the breeding adults at each nest in 2005 and 
2007. Two units (#\s 85753 and 85754 [North Star 
Science and Technology. King George. VA. 
USA]), providing locations every 3 hrs. were 
attached to a pair of adults at a single nest in 
2009. All six transmitters had an accuracy of 
- 10-15 m. Transmitters for four adult vultures 
recorded the daily maximum elevation above 
ground and the time when this occurred. We used 
DNA from feathers to identify individuals as male 
or female (Health Gene Laboratories, Toronto, 
ON. Canada) for a pair of adults captured at the 
same nest site (male # 85753 and female # 
85754), 
Breeding Home Ranges.— The 2005 and 2007 
migration paths to Venezuela and back appear in 
Mandel et al. (2011). We used locations after 
birds returned to Saskatchewan (Fig. 1) during 
weeks when adults were incubating eggs or 
feeding young (Jun-Aug) to calculate home range 
size. We used 95% Minimum Convex Polygon 
(MCP: While and Garrotl 1990) and 95% fixed- 
kernel (Worton 1989, 1995) methods, using the 
Home Range Estimator < HRE) program of Rogers 
and Carr (1998) available for AreView 3.2 (ESR1 
2005). The 95% kernel contour is believed to 
most accurately reflect home range size because it 
minimizes biases caused by inclusion of outlying 
locations in the range estimate (Kemohan et al. 
2001, Hasselblad and Bcchard 2007). GPS 
telemetry data generated large data sets, including 
multiple locations (up to 24) on single days. We 
re-sampled all data sets to 120 randomly chosen 
points to reduce autocorrelation for computing 
fixed-kernel home ranges, which are utilization 
distributions of use sensitive to temporal autocor¬ 
relation (Schoener 1981; Swihart and Slade 
1985a. b; Ackerman et al. 1990). We computed 
the smoothing parameter (h) of fixed-kernel 
ranges as the optimum value with reference to a 
known standard distribution (i.e.. Silverman 1986; 
Worton 1989, 1995). This was the square root of 
the mean variance in x and y coordinates divided 
by the sixth root of the number of points following 
a standard bivariate normal probability density 
function (Rogers and Carr 1998). We used this 
method because, after 1 week of brooding, each 
adult vulture typically visited the nest site once 
daily, and /?-ref is expected to be effective if the 
underlying use distribution is unimodal (Worton 
1995. Rogers and Carr 1998). Statistical analyses 
were performed in R Version 2.10.0 (R Founda¬ 
tion for Statistical Computing 2009). Data pre¬ 
sented are means ± SD. 
RESULTS 
We obtained an average of 988 ± 392 satellite¬ 
tracking locations per bird per year (Table 1). 
Home range size estimated using the 95% MCP 
method ranged from 47 to 953 km 2 , averaging 371 
± 340 km 2 (Table 1). Fixed-kernel home ranges 
(95%) ranged from 49 to 1,992 km 2 and averaged 
648 ± 731 km 2 (Table l). Fixed-kernel ranges 
calculated from 120 randomly selected points to 
minimize temporal autocorrelation in each telem¬ 
etry location did not differ from fixed-kernel 
home ranges calculated using all points available 
for a bird (paired two-sample /-test, t = -1.32, P 
= 0.23). Mean home-range size calculated using 
the 95% MCP and 95% fixed-kernel methods also 
did not differ (paired two-sample /-test, t = -0.28, 
