Saborse ami Renne • HISTORIC PRESENCE AND ABSENCE OF PREDATORS 
751 
(Trumbull. Mahoning, and Columbiana) from 
February through early May 201 1. including 
locations in Poland Township, Yellow Creek 
Park. Mosquito Creek State Park, Mill Creek 
Metro Parks. Berlin Lake, and Beaver Creek State 
Park (Fig. 1). We captured two instead of one 
chickadee in three locations without and one 
location with screech-owls; capture sites ranged 
from 1.6 to 6 km apart. Two chickadees were 
caught within 100 m of each other at two of our 
10 sites; the time between these captures was at 
least 6 weeks. The closest capture site with a 
known Eastern Screech-Owl population to an area 
without them was 17 km, a distance greater than 
the screech-owl median natal dispersal distance of 
2.3 km (Belthoff and Ritchison 1989) and average 
circular home range radius of 400 m (Belthoff 
et al. 1993). This greatly reduced the chance that 
captured chickadees had not dispersed between 
areas with and without screech-owls, particularly 
since the median natal dispersal distance for 
chickadees is l.l km (Weise and Meyer 1979); 
their average circular home range radius in w inter 
is 120 to 230 m (Smith and Buskirk 1988). 
Chickadees were caught using a Sparrow Sled 
Live Trap (Wildlife Control Supplies. East 
Granby, CT. USA) which was hung near a 
feeding station with a 1.8-m Sheppard’s hook. 
The trap was covered with a sheet after an 
individual was captured and the chickadee was 
allowed to calm dowm for 5 min. Nearby I’lock 
members can alter calling behavior (Frceberg and 
Haney 2008) and this time was extended if Con¬ 
or heterospecific Hock members were within a 
30-ni radius of the trap or could be heard. All 
individuals were tested in isolation. A Sennheiscr 
ME67 shotgun microphone w-as placed on the 
■fap. the sheet was lifted and a stuffed Eastern 
Screech-Owl mounted within 0.5 m of the trap 
was revealed (the owl model, in a perched 
position, was loaned from the Cleveland Museum 
of Natural History by A. W. Jones). Calls pro¬ 
duced by the individual were recorded tor lip to 
10 min using a Maranlz professional portable 
s °lid state recorder (model # PMD670) and the 
chickadee was released. We ceased recording if 
an individual stopped calling tor 5 min because 
we felt it had habituated to the model. 
Call and Statistical Analysis. —‘C!iick-a-dee 
calls were analyzed using Raven Lite software 
(Cornell Laboratory of Ornithology. Ithaca. NY, 
USA). We recorded for each individual: (1) total 
number of calls, (2) average number of D notes 
per call. (3) average duration of the first D note, 
(4) average length of time between the First and 
second D note, and (5) average length of time 
(see) between the ‘chick-a’ and D note. These 
‘chick-a-dee’ structural components are known to 
vary based on predator threat level (Templeton el 
al. 2005, Soard and Ritchison 2009). We be¬ 
lieved a priori the acoustic structure initially 
emitted would best represent the information 
conveyed for that predator and only analyzed 
components of the first five ‘chick-a-dee’ calls 
(i.e.. before habituation may have occurred). We 
caught chickadees across large spatial and tem¬ 
poral scales and treated each captured individual 
as an independent experimental unit. Four and 
eight individuals were respectively caught from 
areas with and without screech-owls. 
We initially used multivariate analysis of 
variance (MANOVA) (Scheincr 1993) to assess 
if historic screech-owl presence or absence af¬ 
fected response variables. All significant main 
effects from the MANOVA were analyzed using a 
one-way ANOVA in SPSS (2009) with screech- 
ow'l presence and absence as treatment levels. 
Treatment variances were heterogeneous for the 
number of calls per individual (P < 0.001, 
Levine’s HOV test) and we analyzed this variable 
using log-transformed data, which resulted in 
homogeneous variances (P = 0.63). Homoge¬ 
neous treatment level variances in the ANOVA 
w'ere found for the other four factors (P > 0.05) 
and no other transformations were necessary. An 
a priori a-lcvel w'as set at 0.05 for all statistical 
tests. 
Eastern Screech-Owl Survey. —Screech-owl 
surveys were performed using methods adapted 
from Takats et ai. (2001) to substantiate historic 
reports of their presence or absence at sites in 
which Black-capped Chickadees were captured 
(Fig. 1). These areas w'ere fragmented deciduous 
forests and woodlots of 4 to 260 ha. which is 
deemed good screech-owl habitat (Sparks et al. 
1994). Surveys began 30 min after sundown 
between November 2010 and January 2012, a 
diurnally active time for screech-owls (Deuser 
2011 ), and each of 10 sites were sampled two to 
four times during this interval (we stopped 
playbacks if a response was detected). Eastern 
Screech-Owl ‘trill' and ‘whinny’ calls were taken 
from Peterson Field Guides (1990) (length of call 
sequence: 24 see): this same series was played at 
each site using a Foxpro Spitfire digital game call 
(model # SF1). The volume was set so playback 
