Reiter and Andersen • ARCTIC FOXES, LEMMINGS, AND CANADA GEESE 
271 
o 
d 
1994 1996 1998 2000 2002 2004 
Year 
FIG. 2. (A) Estimated Eastern Prairie Population Canada Goose probability of nest success. (B) Canada Goose nest 
dens 'ty <nests/100 ha wetland nest habitat). (C) Median hatch date (relative to 1 Jun = 1). (D) Index of collared lemming 
Sundance based on willow scar analysis (expected number ol scars). (E) The proportion ol occupied arctic lox dens. Error 
present 95% confidence intervals for all estimates. 
!n tiie mid-1990s but has become more variable in 
recent years (Fig. 2B). The autocorrelation func- 
1,011 provided no evidence of regular cycles 
fetween 1993 and 2004 (Fig. 3B). Median hatch 
^ ranged from 13 June to 11 July and averaged 
-6 June. Median hatch date has become more 
triable in recent years (Fig. 2C) but exhibited no 
s '£nificant evidence of regular cycles (Fig. 3C). 
,VJest density and median hatch date were 
negaiively correlated ( rho = -0.49. P = 0.10), 
88 years with late hatch dates tended to have lower 
density. 
identified five peak lemming population 
years (1995, 1998, 2000, 2001, 2003) and four 
lemming trough years, 1 year following a peak 
(1993. 1996, 1999, 2002) between 1993 and 2004. 
This was based on our index to relative lemming 
abundance (i.e.. predicted number of willow scar- 
age counts per year) which ranged from 72 to 477 
with a mean of 279 ± 150. Relative lemming 
abundance cycled with a periodicity of 3 years at 
Nestor One during 1993-2004 (Fig. 2D), and the 
autocorrelation function identified significant 
positive correlation in lemming abundance at 3- 
year intervals between 1993 and 2004 (Fig. 3D). 
The proportion of occupied fox dens ranged 
from 0.13 to 0.91 between 1993 and 2004 and 
averaged 0.61 ± 0.25. The total number of dens 
