376 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 124, No. 2, June 2012 
TABLE I. Estimates of Hermit Thrush territory densities (some estimates were converted to density/ha to facilitate 
comparison with current study). 
Geographic location 
Density (per ha) 
Source 
Wisconsin 
0.1-3.1 
Direct observation of nests 
Arizona 
1.9 
Spot mapping 
Wisconsin 
0.8 
Direct observation of nests 
Nova Scotia 
0.6 
Spot mapping 
Colorado 
0.2-0.3 
Spot mapping 
Wyoming 
<0.1 
Strip transect counts 
California 
<0.1 
Spot mapping 
Michigan 
<0.1 
Direct observation of nests 
Study 
Current 
Franzreh and Ohmart 1978 
Flaspohler et al. 2001 
Morgan and Freedman 1986 
Snyder 1950 
Salt 1957 
Bock and Lynch 1970 
Pettingill 1930 
invaded by Lumbricus earthworms (average 
biomass = 5.74 g/m), and three were effectively 
earthworm-free (i.e., either containing no earth¬ 
worms or only Dendrobaena octaedra , a small 
species with comparatively minor effects on the 
forest floor) (average biomass = 0.05 g/m'). 
Follow-up sampling at all sites in 2008 indicated 
they had not changed invasion status (Loss and 
Blair 2011). Four of the original 10 sites were not 
used; two had variable invasion status within the 
site and two had only one Hermit Thrush nest 
found in each year. Sites were 10-20 ha in size, at 
least 2 km from each other, and all were in sugar 
maple ( Acer sacchartmi)/ basswood (Tilia ameri- 
cana) forest stands >60 years of age that had 
experienced no timber management in the last 
40 years and were on sandy-loam or loamy-sand 
soils (Holdswonh et al. 2007a). Remnant patches 
of eanhworm-free forest in the study region are 
far from roads, fishing locations (e.g., boat 
launches and lakeshores). and human settlements 
(Holdsworth et al. 2007b). The earthworm-free 
sites in the current study had likely escaped 
invasion because of their remote locations in 
designated wilderness areas that were >1.6 km 
from the nearest road. 
I searched for and monitored Hermit Thrush 
nests between 20 May and 14 July 2009 and 17 
May and 9 July 2010, taking care to expend 
equivalent search effort at all sites and to evenly 
distribute search effort within each site. I used a 
geographical positioning system (C.PS) to record 
all nest locations and monitored nests every 3- 
4 days until they had successfully Hedged at least 
one nestling, or were depredated or abandoned. 
Nests were classified as successful if empty after 
the expected Hedging date. (I M2 days: Jones and 
onovan 1996), an d I observed fledglings 
agitated adults, or droppings near the nest. Nests 
were classified as depredated if eggs were 
destroyed or if nestlings were absent from ihe 
nest before the expected Hedging date with no 
evidence of Hedging. Nests were classified as 
abandoned if eggs remained in the nest with no 
adult activity during four consecutive visits. 
1 used Arc Map Version 9.3 (ESRI 2008) lo 
calculate a matrix of distances among (I) all nests 
in a 1.3-ha valley that was part of one earthworm- 
Irec study site (but spatially separated from the rest 
of the site by a 15—25 m high ridge system), and 
had a high density of Hermit Thrushes and nests 
(hereafter Porcupine Lake) (Fig. i); and (2) all 
nests (excluding Porcupine Lake) within each of 
the six study sites. All distance values within each 
site were averaged to calculate mean distance 
among nests. Nest density was calculated tnests 
divided by area searched) for Porcupine Lake and 
lor each site (excluding Porcupine Lake). .Areas 
used corresponded to the perimeter within which I 
searched for nests at each site and the 1.3-ha area of 
Porcupine Lake. All calculations were conducted 
separately for 2009 and 2010; however, data were 
averaged across years for both calculations. All 
density estimates are reported as nests per hectare 
Independent samples /-tests were used to 
compare nest density and mean distance among 
nests between Lumbricus-f ree and invaded sites. 
Program MARK'S nest survival package (Dins- 
more and Dinsmore 2007). w hich accounts for the 
exposure (i.e.. number of days observed active) ol 
each nest, was used to separately estimate nest 
success probability for nests at Porcupine Lake 
(« = 8) and for all other nests (n = 64). The 
sample size of nests at Porcupine Lake was small 
and 1 also report observations on whether 
individual nests were successful. 
