Reiter and Andersen • ARCTIC FOXES, LEMMINGS, AND CANADA GEESE 
273 
•1.0 -0.8 -06 -04 -02 00 
Fox den coefficient 
FIG. 4. Frequency of 800 coefficient estimates for the 
association of the proportion of occupied arctic fox dens 
IFD) with EPP Canada Goose nest success from 200 
random iterations of arctic fox den data and evaluation of 
four models with FD included. 
(■ 7 0/173). The second-best-supported model from 
tile original analysis was the only other to be 
selected as the top model (39% of iterations) and 
included the fox trap effect; another index to fox 
abundance. Coefficient estimates for fox dens 
(Pro; n = 800) from our simulation were not >0 
and the 95% Cl was -0.61 to -0.39 (Fig. 4). 
TTicsc diagnostic simulations suggest that uncer¬ 
tainly in our estimates of the proportion of 
occupied fox dens in each year did not signifi¬ 
cantly influence model selection or inference 
regarding the influence of fox den occupancy on 
Canada Goose nest success. 
DISCUSSION 
Nest success of dispersed-nesting Canada 
^ eese in our study area within a subarctic 
coastal-tundra trophic ecosystem near Cape 
Churchill. Manitoba, exhibited substantial year- 
to-year variation in the mid 1990s and early 
-000s. Arctic fox den occupancy and nest density 
Were both significantly associated with Canada 
Goose nest success. Our best supported model of 
'-anada Goose nest success, however, was not 
consistent with predictions of the bird-lemming 
hypothesis, as higher arctic fox den occupancy 
lowered nest success and was not related to 
lornining abundance. Walter (1999) reported that 
'Creased nest depredation by arctic fox was the 
primary cause of low nest success at Nestor One 
in the 1980s. Our assessment of the effects of 
arctic fox supports the contention that foxes are a 
primary predator of nests in this system and, after 
spring phenology (presumably as it influences 
nest density, hatch date, and the condition of 
nesting geese), arctic fox are an important factor 
influencing Canada Goose nest success. The 
mechanisms driving changes in the fox population 
in this ecosystem, which we were unable to 
directly link with lemming abundance and the 
bird-lemming hypothesis, have important conse¬ 
quences for nesting Canada Geese and likely other 
birds that nest along western Hudson Bay. 
Arctic fox den occupancy varied among years, 
although non-cyclically. This is consistent with 
reports of variation in arctic fox reproductive 
behavior elsewhere where primary food resources 
fluctuate (Macpherson 1969, Elmhagen et al. 
2000, Strand et al. 2000). Lemming abundance 
fluctuated cyclically (as postulated under the bird¬ 
lemming hypothesis) on our study area, but was 
not included in our best-supported model of 
Canada Goose nest success. We were unable to 
identify an association between arctic fox den 
occupancy and peak lemming years as predicted 
by the bird-lemming hypothesis. Our results were 
consistent with those of Bromley et al. (1998) who 
found reduced nest survival of Cackling Geese 
(Branta hutchinsii) during high fox years in the 
Northwest Territories, Canada, and with Reiter 
(2009) who identified high den occupancy in high 
lemming years and lower Canada Goose nest 
survival at Cape Churchill (2005-2008). Our data 
also suggest arctic fox removal was effective at 
increasing Canada Goose nest success, supporting 
previous research that identified arctic fox as a 
critical component influencing nest success in this 
system (Walter 1996, 1999). 
Our results, similar to Bety et al. (2001) and 
Miller et al. (2007). indicate density of nesting 
geese was positively associated with nest success. 
This positive association was cither the result of 
(1) early spring phenology and nesting geese in 
good physiological condition, which allowed for 
more time to tend nests and more effective 
defense of nests against predation; (2) a higher 
density of goose nests that induced group defense 
against arctic fox; or (3) a combination of these 
two factors. The evolution of dispersed nesting 
ecology by Canada Geese may be the result of the 
individual defensive capabilities of adult Canada 
Geese; a nesting pair often can successfully 
defend their nest against arctic fox (Bahr 1989). 
Group defense has rarely been observed in nesting 
