Red-cockaded Woodpecker Home Range 39 



territories of other family groups, because the amount of time spent 

 foraging within a territory and the time spent in territorial maintenance 

 may depend upon territory quality (Ewald and Carpenter 1978). 



A set of vegetation sampling points was selected from each overall 

 home range by cluster sampling. Parallel transects traversing the home 

 range were located perpendicular to an axis that was nearly parallel to 

 most streams. Transects were located randomly along the axis but could 

 not be less than 60 m apart. A sample point was located randomly 

 within each 90-m segment of a transect. Distances were paced along a 

 compass bearing. One hundred sample points were selected for group A, 

 and 49 were selected for group B. 



A random sample of foraging areas was selected from each tracking 

 period. Foraging areas were defined as locations in the tracking itinerary 

 in which the scan sample of behavior indicated that the family group 

 was foraging. We randomly chose 25 foraging locations from each 

 tracking period, except the final one for which 24 were chosen because 

 of an error in a computer program. The samples averaged 5% of the 

 foraging locations within tracking bouts. Each location was relocated in 

 the field, and a random distance of up to 23 m (75 feet) was paced in a 

 random compass direction to offset potential investigator bias in 

 relocating points. As a further safeguard, distance and direction of each 

 deviation were not ascertained until the point was relocated. 



Each sampling point was the center of a Bitterlich variable-radius 

 sampling plot (Husch et al. 1982) defined with a ten-factor prism. This 

 method effectively samples trees of different sizes with plot sizes most 

 suitable for them. For example, trees of 10 cm diameter at breast height 

 (DBH) are sampled with a plot size of 33 m 2 , whereas trees of 45 cm 

 DBH are sampled with a plot size of 691 m 2 . Species; DBH, rounded to 

 the nearest 0.25 cm (= 0.1 inch); and tree height, rounded to the nearest 

 0.3 m (= 1 foot) were recorded for each tree. Pine stems less than 2.5 cm 

 DBH were excluded from the analysis because the birds were not 

 observed foraging on them. Hardwoods were considered understory 

 trees if they were shorter than the mean pine height. Bole surface area 

 was calculated as the surface area of a cone with base on the ground, 

 apex at the tree height, and diameter (DBH) at breast height. For each 

 plot, pine bole surface density (m 2 /ha) and tree density (trees/ ha) were 

 calculated (Husch et al. 1982). Means of bole surface per tree (m 2 /tree), 

 pine DBH (cm), pine height (m), and understory hardwood height (m) 

 were calculated as weighted means using the density expansion factor 

 for each tree as its weight. 



We used the Wilcoxon two-sample test (Sokal and Rohlf 1981) for 

 all comparisons. Rank tests are recommended in resource-use studies 

 because of the imprecision with which resource availability is measured 



