314 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 2. June 2012 
TABLE 3. Regression coefficients on a logit scale for the minimum model exploring the relationship between daily 
survival rate and distance to habitat edges for Sprague's Pipit at Bowdoin National Wildlife Refuge, Montana using 
Akaike’s Information Criterion, corrected for small sample sizes. T = date of nest initiation. Age = age of nest. 
MIN(Edges) = minimum distance to linear features. 
(V-estimale 
SB 
Lower Cl 
Upper a 
{S[T + Age + Clutch Size + MINI Edges)]) 
Intercept 1.62 
0.95 
-0.25 
3.49 
T 
0.01 
0.01 
0.00 
0.02 
Age 
-0.05 
0.02 
-0.09 
-0.01 
Clutch Size 
0.32 
0.18 
-0.04 
0.68 
MINI Edges) 
0.00 
0.00 
0.00 
0.00 
not improve the base model containing T. age. and 
clutch size (Table 2). We considered different 
distance variables to evaluate the effects of roads 
and other edges, given our minimum AIC,. model 
with T. age, and clutch size. A model with five 
additional covariates lor the five distance catego¬ 
ries was considered, but was poorly supported 
(Table 2). likely because of the excessive number 
ol parameters allocated to the distance portion of 
the model. We used the minimum ol' the distances 
to roads (MINIRoads]) and the minimum of the 
distances to any edge as predictive variables. We 
also considered models with thresholds (within 
50. 100, 200. or 300 m) on these minimum dis¬ 
tance variables; we would not expect the effect of 
distance to edge to be important beyond these 
threshold values. None of the distance models 
improved the minimum AIC c model with T, age 
of nest, and clutch size (Table 2). There was no 
support for variables with distance to any of the 
edges, including to roads, having an effect on 
DSR relative to the minimum AIC,. model that 
contained three non-distance covariates (Table 3). 
DISCUSSION 
We observed no relationship between nest 
survival and distance to roads, or to habitat 
changes and edges at Bowdoin NWR. We 
controlled (or time-specific factors including age 
and date of clutch initiation (season) and found 
these to be more important in explaining DSR 
than distance to edges, similar to results in 
Saskatchewan (Davis ct al. 2006), However, the 
difference (AAIC,.; Table I) between the best and 
the worst models was small, suggesting that no 
model explained a large portion of the variation 
in nesting success. Nest depredation typically is 
the primary cause of nest failure in grasslands 
(Johnson and Temple 1990, Vickery et al. 1992. 
Martin 1993, Davis 2003. Jones and Dieni 2007). 
Some predators may use roads for hunting and 
travel; however, the predator community in the 
area studied is diverse and opportunistic (Jones 
et al. 2010), and predation could occur indepen 
dent of proximity of the nest to edges. 
Differences among studies that have evaluated 
an effect of distance to roads on nest survivorship 
could reflect several confounding factors. Many 
studies that have evaluated road effects have 
focused on primary roads (Fahrig and Rytwinski 
2009), which are characterized by greater vehic- 
ular traffic, greater landscape alterations (e.g., 
buildings, suburban development), and othei 
disturbances (e.g., oil, gas. wind energy develop¬ 
ment; Dale et al. 2009) that are not associated 
with secondary and tertiary roads and trails 
Higher disturbance levels on primary roads may 
not he comparable to those in rural northern 
mixed-grass prairies. Bowdoin NWR is crossed 
by many diked vehicle roads and trails; however, 
these roads generally have low and seasonal 
traffic levels. Bowdoin NWR occurs within •' 
landscape expanse of native mixed-grass praine 
Grasslands in this region are characterized bv e 
low proportion of edges and a high interior a/ea- 
to-edge ratio, which may be the reason for the 
lack of effects on nest survivorship in our stud; 
Our work at Bowdoin NWR represents only 
one point in both geography and time; however 
there arc few studies on the effects of roads and 
edges on Sprague's Pipits (Sutter el al. 30<" 
Davis 2004. Koper and Schmiegelow 2006. Koper 
et al. 2009). The response of Sprague's Pipits 
may vary geographically, similar to area sensitivity 
in other grassland birds (Johnson and Igl -2001 1 
Conservation of Sprague's Pipits may require 
information on relative levels of the threats to thei' 
populations throughout their range. Combining data 
