Content uploaded by Jerzy Romanowski
Author content
All content in this area was uploaded by Jerzy Romanowski
Content may be subject to copyright.
The Auk 114(4):628-637, 1997
ROOSTING BEHAVIOR AND GROUP TERRITORIALITY IN
AMERICAN CROWS
DONALD E CACCAMISE, • LISA g. REED, JERZY ROMANOWSKI, 2 AND PHILIP C. STOUFFER 3
Department of Entomology, Rutgers University, New Brunswick, New Jersey 08903, USA
ABSTRACT.--Cooperative groups of American Crows (Corvus brachyrhynchos) maintain
group territories year-round while often traveling long distances to roost communally at
night. Our goal was to discover how territorial crows resolve the conflict between the co-
hesive nature of group behavior with requirements of dispersal to roost communally. We
color-marked crows to study group composition over two years, and radio-tagged crows to
study movement among roosts, territories, and feeding areas. Most crows showed diurnal
fidelity to the group territory throughout the year. Yet, most birds frequently left territories
during the day to forage up to 4 km away. At night, crows roosted either on their territory
or 18 km away at a large roost adjacent to a landfill. Crows roosted on territories more often
in spring (87%) than in winter (42%). Group cohesion was high on territories, yet we found
no evidence for group behavior away from territories. Crows arrived singly both to terri-
tories in the morning and to the communal roost in the afternoon. Group cohesion for ter-
ritorial crows appears to be based on decisions of individuals to return to territories from
distant roosting and foraging sites. Group cohesion on territories is tied to retention of
breeding sites, whereas dispersal for communal roosting likely is linked to benefits derived
from foraging away from territories, particularly in winter when physiological stress is
greatest and territorial food supplies are lowest. Received 10 May 1996, accepted 9 April 1997.
COOPERATIVE BREEDING AND COMMUNAL
ROOSTING in large aggregations have attracted
considerable recent attention (e.g. Allen and
Young 1982, Brown 1987, Caccamise 1993). Co-
operative breeding and communal roosting are
forms of group behavior, yet we know of no
studies that have simultaneously examined re-
lationships between these behaviors. Occur-
rence of both behaviors in a single species ap-
parently is uncommon and may explain the
lack of attention from researchers. American
Crows (Corvus brachyrhynchos) offer the oppor-
tunity to examine relationships between coop-
erative breeding (Chamberlain-Auger et al.
1990) and communal roosting (Stauffer and
Caccamise 1991a) because crows frequently
travel to large communal roosts, yet they live
in stable groups that breed caaperatively and
defend year-round territories.
In the northeastern United States, large cam-
• E-mail: caccamis@rci.rutgers.edu
2 Present address: Department of Vertebrate Ecol-
ogy, Institute of Ecology, Polish Academy of Science,
05-092 Lomianki, n. Warsaw, Dziekanow Lesny, Po-
land.
3 Present address: Department of Biological Sci-
ences, Southeastern Louisiana University, Ham-
mond, Louisiana 70402, USA.
munal roosts of American Crows form each
evening from late summer through early
spring. The largest aggregations develop in
winter (Goodwin 1976). Cooperative breeding
and communal roosting have what appear to
be conflicting requirements in that group ter-
ritorial behavior requires group cohesion on
territories, whereas communal roosting re-
quires travel to distant sites with long absences
from territories. It is not known how stable
groups occupying territories during the day
are related to large assemblages that form com-
munal roosts at night. Nor is it known why in-
dividual crows that belong to stable groups
abandon territories to join distant aggregations
at communal roosts.
Our goal was to understand how crows re-
solve apparent conflicts between the group co-
hesion required for cooperative breeding and
territory defense with the temporary abandon-
ment of territories necessary to join communal
roosts. We used color-marking to study group
composition and radiotelemetry to study
movements of individuals among roosts, ter-
ritories, and feeding areas.
METHODS
Our study area was centered on the Cook College
campus of Rutgers University, New Brunswick, New
628
October 1997] Roosting and Territoriality in Crows 629
Jersey. It included the open areas on the southern
part of campus and adjacent agricultural fields
(hereafter "Cook"). We also worked at Edgeboro
Landfill and Middlesex County Yard and Vegetative
Waste Cornposting Facility located about 4 km east
of Cook. A third study area was at a communal roost
located 18 km ENE of Cook on Staten Island.
We captured crows either with the sedative alpha-
chloralose (Stouffer and Caccamise 1991b) or with a
WinnStarr rocket net. We used sedatives to capture
only one crow because rocket nets were more reliable
(22 crows). We attached transmitters (L. L. Electron-
ics) to all captured crows using a backpack (Stouffer
and Caccamise 1991a). Transmitters were two-stage
(16 to 19 g) or one-stage (9 to 12 g) that weighed
-<3.5% or about 2% of each crow's body mass, re-
spectively. We also marked birds individually using
hair stripper (Romanowski et al. 1993) to bleach
unique patterns on the wing and tail feathers.
In 1990-91, we captured, marked, and radio-
tagged 12 crows from five territories. Crows were
aged using mouth and covert coloration (Pyle et al.
1987). One crow disappeared the day after release
and was seen only periodically during the next two
months. It was considered a vagrant and was not in-
cluded in the analysis (Stouffer and Caccamise
1991a). The remaining 11 crows included five adults
and six subadults (to second year) from five group
territories. We banded seven nestlings from nests in
two of these territories.
In 1991-92, we captured, marked, and radio-
tagged 10 crows from seven territories, including
eight adults and two subadults. Two adults were ra-
dio-tagged the previous year, and one juvenile had
been banded as a nestling the previous year. In ad-
dition to radio-tagged birds, we identified six indi-
viduals with unique physical characteristics that al-
lowed visual identification. These included five pre-
viously marked birds and one crow that consistently
had a drooping wing. We banded 29 nestlings from
10 nests in 1992. Five nests were located in territories
with crows that were radio-tagged the previous win-
ter.
In 1990-91 (1 November 1990 to 7 October 1991),
we attempted to locate each crow once a day, six days
a week, on a fixed schedule such that each bird was
located twice between 0800 and 0930, twice between
0930 and 1100, and twice between 1100 and 1230. In
1991-92 (20 December 1991 to 1 December 1992), we
attempted to locate crows once a day, six times a
week. We blocked times into five periods (0800 to
1000, 1000 to 1200, 1200 to 1400, 1400 to 1600, and
1600 to 1800) and sampled each period sequentially
on successive days. This resulted in approximately
equal samples for each time interval. In both years
when we could not locate birds on their territories,
we searched for approximately one hour before con-
sidering them absent. Search areas included all sites
at which radio-tagged crows had occurred previous-
ly during the day as well as novel sites that we con-
sidered potentially attractive to crows.
Our telemetry studies were divided into one
spring and two winter seasons. We defined the be-
ginning of the winter season as the date we initially
released radio-tagged birds (1 November 1990 and
20 December 1991). The winter season extended un-
til we first discovered an incubating female (9 April
in 1991 and 1992). We continued spring observations
through the field lives of individual transmitters (12
July to 22 September).
In both years, we located all birds roosting on their
territories six nights a week. In 1990-91, this includ-
ed determining the exact trees occupied by the birds
and the number and identity of roost mates. In 1990-
91, we located the radio-tagged crows at their com-
munal roost on Staten Island four to seven nights a
week, and in 1992 we sampled the Staten Island roost
once a week during spring, and once every other
week during summer On roost visits we determined
direction and time of arrival of radio-tagged birds.
We determined the time and direction of departure
of tagged birds from the distant roost at dawn once
or twice a week. We also determined the time birds
left from or arrived at their territories.
RESULTS
Diurnal activity in relation to group behavior.-
Over the two years of study we attempted to
locate our 21 radio-tagged birds 2,208 times
during daylight hours (Table 1). On most at-
tempts we found the birds within the same
small area. We refer to these areas where in-
dividuals were reliably located during the day
as the diurnal activity center (DAC). DACs
were characteristic for individual birds, but
when combined for group members (see be-
low) they formed the territory for the cooper-
ative group. We used the rate of our success at
locating individuals on their DACs as a mea-
sure of fidelity. DAC fidelity ranged from 54 to
96% (œ = 84 + SD of 10.2%) and did not differ
between adults and juveniles (arcsine transfor-
mation; t = -0.88, df = 19, P > 0.39).
Radio-tagged crows were absent on 16% (350
of 2,208) of our scheduled field surveys. For
31% (110 of 350) of these absences, we located
the birds at a landfill (Edgeboro Municipal
Landfill) used by many crows from the area.
Our radio-tagged crows were seen an addition-
al 139 times at the landfill during other phases
of our study. We found all but two of our radio-
tagged crows at this site at least once. We as-
sociated the 11 radio-tagged birds with five dif-
ferent groups in 1991 and the 10 radio-tagged
630 CACCAMISE ET AL.
TABLE 1. Fidelity to diurnal activity centers (DAC) by radio-tagged American Crows.
[Auk, Vol. 114
Success-
Diurnal ful
observa- diurnal Fidelity
Bird Date Date last tions observa- to DAC % of time Group
number Age a radio-tagged observed attempted tions (%)b found c membership
115 AHY 10 Dec 90 21 May 91 94 90 95.7 97.9 TF 91
117 HY 10 Dec 90 30 May 91 97 76 78.4 92.8 TF 91
118 HY 10 Dec 90 7 Oct 91 155 129 83.2 87.8 TF 91
119 AHY 23 Dec 90 14 Jun 91 89 77 86.5 86.5 PG 91
120 AHY 23 Dec 90 30 Apr 91 83 78 94.0 96.4 PG 91
121 HY 24 Dec 90 19 Jun 91 94 79 84.0 84.0 FS 91
122 HY 24 Dec 90 12 Jun 91 92 68 73.9 81.5 FS 91
123 AHY 7 Jan 91 28 Jun 91 96 87 90.6 90.6 PA 91
124 AHY 7 Jan 91 29 Jul 91 112 103 92.0 93.8 PA 91
125 HY 28 Jan 91 22 Apr 91 57 47 82.5 82.5 HB 91
126 HY 28 Jan 91 17 Jul 91 90 77 85.6 87.8 HB 91
117 AHY 18 Dec 91 27 May 92 96 65 67.7 80.2 TF 92
120 AHY 8 Apr 92 9 Oct 92 127 107 84.3 94.5 PG 92
130 HY 31 Dec 91 7 Oct 92 132 121 91.7 92.4 PA 92
135 HY 10 Feb 92 13 Oct 92 172 164 95.3 95.9 GB 92
137 AHY 2 Jan 92 4 Aug 92 155 134 68.5 91.0 FS 92
138 AHY 31 Dec 91 30 May 92 105 57 54.3 68.6 PA 92
139 AHY 20 Jan 92 30 May 92 87 64 73.6 83.9 RT 92
141 AHY 10 Feb 92 18 Jul 92 116 94 81.0 92.2 GB 92
143 AHY 3 Mar 92 22 Jul 92 101 95 94.1 95.0 PO 92
144 AHY 3 Mar 92 9 Jun 92 58 46 79.3 79.3 PO 92
HY, hatching year; AHY, after hatching-year.
Percent of total attempts successful in locating radio-tagged bird on its territory.
Includes birds absent from territory but located at distant sites.
birds with seven groups in 1992. Additionally,
we were aided in evaluating group composi-
tion by having six birds with unique physical
characteristics (i.e. our color marks or physical
attributes) associated with four different
groups. We have seen these unique individuals
in the same groups for as long as three years.
Each time we located a radio-tagged crow,
we counted the crows in its group and identi-
fied any unique individuals. Counts of group
size often varied on our successive encounters
(Table 2). We estimated the number of birds in
each group based on several factors including:
(1) our analysis of group-size frequencies, (2)
observations of unique individuals, and (3) our
familiarity with patterns of group movements
and spatial organization. Our estimates of
group size ranged from three to seven individ-
uals (Table 2).
We used overlapping spatial distribution of
DACs for individual birds in combination with
our familiarity with associative behavior of
group members to deduce boundaries of group
territories (Fig. 1). Group territories were co-
operatively defended throughout the year (Kil-
ham 1989). Conflict between neighboring
groups varied seasonally but was never com-
mon. Aggression was minimal except in late
winter and spring, when we occasionally saw
vocalization and patrol flights, particularly in
early mornings.
We have evidence from nine individuals that
crows remained on the same territories in suc-
cessive years. In addition to the six crows with
unique marks, we recaptured three birds in
two consecutive years, each on the same terri-
tory in which they were originally caught. One
bird initially was captured as an adult, the sec-
ond as a hatching-year bird, and the third was
banded as a nestling and then captured and ra-
dio-tagged the following winter.
Roosting behavior.--Radio-tagged crows roost-
ed either locally on their territories or at a large
roost 18 km from our study area (Fig. 1). Ag-
gregations at local roosts always were small,
usually from one to four birds. At most, only
one or two individuals in a local roost were ra-
dio-tagged, but occasionally we counted up to
four birds when we witnessed the arrival of
crows or when the crows were flushed before
they had settled in for the night.
Crows roosting away from their territories
October 1997] Roosting and Territoriality in Crows
TABLE 2. Frequency of group sizes for cooperative groups of American Crows.
631
Group size Estim.
Year Group 1 2 3 4 5 6 7 >7 size a
1991 TF 15 47 30 13 6 1 0 0 4
1992 TF 9 12 16 3 0 1 0 0 4
1991 PG 2 22 42 6 1 0 0 0 4
1992 PG 16 14 6 1 2 0 0 0 3
1991 FS 19 9 19 23 17 1 0 0 6
1992 FS 9 21 17 8 5 2 1 2 5
1991 PA 11 12 30 30 5 1 1 2 5
1992 PA 14 25 29 23 12 3 3 4 6
1991 HB 7 5 10 9 15 11 1 1 7
1992 GB 22 15 19 17 6 0 1 0 5
1992 PO 11 10 16 9 3 2 0 3 5
1992 RT 11 6 9 2 5 3 3 3 6
Estimated sizes of groups based on group-size frequencies and interpretation of group dispersion and intergroup interactions.
always flew to the same large roost on the
southern end of Staten Island. We estimated the
size of this winter roost at 10,000 to 15,000
crows. In winter, crows left their territories be-
ginning in mid-afternoon (• = 1452 + SE of
11.4 rain, n = 38, range 4 h 47 rain), but gen-
erally they arrived at the communal roost with-
in an hour of sunset (• = 1653 + 4 rain, n = 95,
range 3 h 28 rain). Group cohesion ended once
the birds left their territories. Individuals from
a group often left the territory at about the
same time in the afternoon. The average sepa-
ration time (interval between departures or ar-
rivals of radio-tagged crows of the same group)
for territory departures was 7.1 - 4.42 rain (n
= 8, mode = 0), but the arrival interval at the
distant roost was nearly twice as long (15.2 -
2.6 rain, n = 27, mode = 1). Separation times
448200•
448150(
448100(
448050(
448000(
447950(
447900C
546000
Cook Colleg e •2•• o
PO group •'•PG group
.T gro? group
547000 548000 549000 550000
UTM Coordinotes (meters)
Edison
Landfill
Landfill
New
0 5
Kilorne[ers
Jersey, Staten
Island
Fresh Kills
< Landfill
Atlantic
Staten
Island
F[C. 1. Map of study area showing point locations of radio-tagged birds in 1991-92 field season (dots);
distant roost on Staten Island, New York (filled circle); and off-territory foraging sites (filled triangles). Areas
enclosed by solid lines in Cook College inset show approximate borders of groups territories based on point
locations and visual observations of marked birds.
632 CACCAMISE ET AL. [Auk, Vol. 114
were significantly longer for roost arrivals than
for territory departures (Mann-Whitney U-test,
U = 55.5, P < 0.038). After members of the
communal roost had settled for the night, we
located the position of radio-tagged crows and
found that crows from the same group roosted
at separate locations.
The roost emptied quickly in the morning,
resulting in similar departure times for group
members (g separation time = 6.1 _+ 3.02 min,
n = 21). However, individual crows probably
returned to their territories independently be-
cause arrival intervals averaged 17.8 -+ 13.02
min (n = 20). Separation times were signifi-
cantly longer for territory arrivals than for
roost departures (U = 83.5, P < 0.001).
We determined seasonal patterns in use of
roosts by locating radio-tagged crows in roosts.
We checked the local roosts more often than the
distant roost because of difficulties in traveling
to the distant site. As a result our data take two
forms: (1) on some nights we confirmed the lo-
cation of each individual by checking both the
distant and local roosts; (2) at other times we
were able to check just the local roosts, allow-
ing us to confirm the location only of those in-
dividuals present. On these nights the location
was unknown for individuals roosting away
from their territories. We report our data as
bird-nights, which represent an observation of
a radio-tagged individual on a single night (i.e.
10 birds on one night would yield 10 bird-
nights). We confirmed the locations of our ra-
dio-tagged birds (type 1 above) for a total of
1,492 bird-nights (775 in winter, 717 in spring).
We failed to locate birds only about 1% of the
time on these nights (0.9% in winter, 1.6% in
spring). These values overestimated actual ab-
sences because some of our failures occurred as
transmitters began to fail. The risk of missing
a bird increased as radio signals became weak
due to aging batteries or damaged antennas.
We recorded 669 bird-nights (572 in winter,
97 in spring) when we could determine only
that radio-tagged birds were absent from their
local roosts (type 2 above). On these nights we
were unable to travel to the distant roost to con-
firm that absences from local roosts indicated
use of the distant roost. Nonetheless, we be-
lieve that these absences represented use of the
distant roost because on at least 99% of the
bird-nights when we confirmed the location of
our birds, they occurred at only two locations,
the roosts on their territories or the distant
roost on Staten Island.
Radio-tagged crows roosted away from their
territories more frequently in winter than in
spring (X 2 = 409.7, df = 1, P < 0.0001). Crows
spent only 42% of bird-nights on their territo-
ries in winter (prior to 10 April), but they used
local roosts 87% of the time during the spring
season. Juveniles roosted locally more often
than adults in winter (51% vs. 36%; X 2 = 30.5,
df = 1, P < 0.001) and during spring (97% vs.
85%; X 2 --- 23.1, df =1, P < 0.001).
Patterns of use for local and distant roosts
varied among individual crows (Fig. 2). Some
birds roosted exclusively at the distant roost
until the end of winter (e.g. crow 117), and oth-
ers alternated between the distant and local
roost (e.g. crow 118). In all cases, use of the dis-
tant roost declined from winter to spring (e.g.
March and April; Fig. 2). By mid-ApriL most
crows either stopped going to the distant roost,
or made individual forays to it after a series of
successive nights at a local roost.
Foraging activity away from the territory.--Most
of the radio-tagged crows left their territories
during the day (Table 1). In nearly all cases, we
located these birds at known foraging sites.
Crows foraged away from their territories in
three circumstances. First, they used foraging
sites near their territories (i.e. within 500 m).
One such site was located at a harvested corn
field adjacent to the southwestern side of our
study area. In fall, abundant post-harvest
waste grain attracted what appeared to be mi-
grant flocks of crows as well as many of our
marked crows from the study area. Second, we
found a foraging site 4 km from our study area
that was used during brief diurnal forays, usu-
ally by individual crows. This site was at the
landfill consisting of a vegetative waste-com-
posting facility and a large municipal landfill
and was in an area of filled marshland adjacent
to the Raritan River (Fig. 1). Third, foraging ar-
eas near or along the way to distant roosts were
used when crows commuted between their ter-
ritories and the Staten Island roost. These sites
were not exclusive of the other two types, but
we considered them distinct because crows
used them after they left their territories in the
afternoon and/or before they arrived at their
territories in the morning. The landfill was the
most frequently used site. In 110 of 125 (88%)
times that we observed radio-tagged crows at
October 1997] Roosting and Territoriality in Crows 633
Staten ' ' ' • ' • 1
Crow 117
Territory • • •
Crow 120
Adult
Staten
Island
Territory
Crow 122
Subadult
Staten
Island
Territory
Staten '• '
Island Crow 118
Subadult
Territory • ,
Jan Feb Mar Apr May Jun
FIG. 2. Seasonal patterns of use of local and Staten Island roosts by radio-tagged American Crows. Solid
triangles indicate confirmed use; open triangles represent occasions when presence of crow at distant roost
was presumed but not confirmed (see text).
the landfill in the afternoon, individual crows
roosted at the distant Staten Island roost. On
only 10 occasions did crows forage at this site
in the afternoon and then return to roost locally
on their territories. Another site frequently
used by our radio-tagged birds was west of the
roost across the Arthur Kill waterway. Diffi-
culties of getting to and working in this area
precluded our successful identification of the
precise site used by the birds. Nonetheless,
commuting radio-tagged crows regularly
stopped at this site as well as at other sites near
the Staten Island roost.
Additional evidence suggests that crows
usually stopped during their commute be-
tween the territory and the distant roost. This
is based on travel times between territories,
distant roosts, and foraging sites. In the morn-
ing, crows took an average of 50 -+ SD of 30.9
rain (n = 43) to travel from the distant roost to
their territories. In the evening, they took more
than twice as long (t = 2.90, df = 10, P = 0.008),
averaging 126 -+ 85.8 min (n = 11) to travel
from their territories to the distant roost.
DISCUSSION
Diurnal movements relative to group behavior.--
Each crow consistently used its own DAC de-
spite frequent trips to a distant roost in the eve-
ning. This pattern occurs in many communally
roosting species (e.g. European Starlings [Stur-
nus vulgaris], Caccamise and Morrison 1988;
American Robins [Turdus migratorius] and
Common Grackles [Quiscalus quiscula], Morri-
son and Caccamise 1990; and Red-winged
634 CACCAMISE ET AL. [Auk, Vol. 114
Blackbirds [Agelaius phoeniceus], Johnson 1979,
Caccamise 1990). For crows, the individual na-
ture of the spatial distributions that form DACs
easily could be overlooked because of the
prominence of their group-based behavior on
territories. Nonetheless, DACs for each of the
radio-tagged crows had unique features, even
among members of the same cooperative
group.
Crows were faithful to their DACs even in
winter when they frequently traveled to distant
roosts at night. The high level of DAC fidelity
is most easily explained by group cohesion.
However, such a clear relationship is not ap-
parent in other roosting species with high lev-
els of DAC fidelity. This is because other roost-
ing species are not known to actively defend
territories outside of the breeding season.
Nonetheless, continued presence in the breed-
ing area probably contributes to long-term oc-
cupancy of a breeding site even when territo-
ries are not maintained in the nonbreeding sea-
son. For example, in winter European Starlings
regularly spend the night in cavities that they
will later use for nesting (Lombardo et al.
1989).
The variation in our counts of group size was
unexpected. We identified three sources: (1)
one or more group members were absent at the
time of the count, (2) group composition actu-
ally changed, or (3) nongroup members were
present on the territory (these were either in-
dividuals from nearby territories or groups of
vagrants). The absence of group members at
the time of field surveys was the most frequent
cause for the variation. Absent birds often were
at nearby foraging sites (see below). In addi-
tion, group membership for some individuals
seemed somewhat labile in that certain crows
were absent for extended periods (i.e. days to
weeks), only to return at varying intervals
without any apparent resistance from the more
stable group members. Similar observations
were made by Caffrey (1992), where dispersion
of juvenile crows ranged from leaving the area
entirely to remaining with the family group.
The only actual changes in group composition
that we saw occurred when groups increased
in size as young of the year took up residency
on the natal territory. Chamberlain-Auger et al.
(1990) concluded that increases in size of crow
groups were due to young remaining on natal
territories. We documented this three times in
our study.
Despite variations in group size, the stability
of groups was evident based on consistent
group composition for the marked crows. Some
individuals remained with the same group on
the same territory throughout the three years of
our study. This stability likely minimized strife
among groups. Overt aggression was rare
among groups, only becoming apparent for
short intervals in March and April.
Roosting behavior.--Radio-tagged crows were
very consistent in their selection of roost sites,
occurring either: (1) alone or in small groups on
their territories, or (2) at the large communal
roost on Staten Island. Nonetheless, the social
behavior related to group membership and ter-
ritoriality differed markedly depending on
where the crows roosted. When roosting on the
territory, crows maintained strong group ties,
often interacting with other group members
throughout the day and up until the time they
settled for the night.
When crows roosted at Staten Island, we
found no evidence for group cohesion once in-
dividuals left the territory. The considerable
differences among group members in the tim-
ing of arrivals at their territories and depar-
tures from the distant roost suggest that de-
parture decisions were made independently.
Similarly, we found no evidence of group co-
hesion when several group members were
present at the distant roost. We detected no co-
ordinated group activity nor any tendency for
group members to roost near one another.
Foraging activity away from the territory.--Ra-
dio-tagged crows frequently left their territo-
ries to forage. Most often this occurred in mid-
to late afternoon when crows left for the day.
However, individuals also left earlier in the day
for intervals ranging from about 10 min to 2 h.
Brief absences (10 to 30 min) generally oc-
curred when crows traveled to foraging sites
away from their territories but within the study
area at Cook College. However, individual
crows sometimes were absent from their terri-
tories for longer intervals (1 to 2 h). These ab-
sences resulted mainly from travel to more dis-
tant foraging sites. We cannot confirm that ev-
ery such absence was a foraging trip, but our
evidence suggests that travel to off-territory
foraging sites accounted for a large proportion
of the absences. Once we identified the forag-
October 1997] Roosting and Territoriality in Crows 635
ing sites used most frequently by radio-tagged
crows, we generally checked these sites when-
ever a crow was missing from its territory. We
located radio-tagged crows at the landfill 31%
of the times they were absent during territory
checks. During the remainder of absences
crows probably used at least one other distant
foraging site. Despite extensive searches within
a radius of 6 to 8 km around the study area,
however, we never located an additional for-
aging site.
The landfill was an important foraging site,
both during the day and in the evening before
crows went to roost. Foraging was the most fre-
quent behavior observed at the landfill, partic-
ularly when crow numbers increased late in the
day. We found that 88% of the radio-tagged
crows that visited the landfill late in the day
roosted at Staten Island that night; only 12% re-
turned to their territories to roost. These ob-
servations suggest that the evening visit to the
landfill was a regular practice of crows before
they undertook the long commute to the roost
at Staten Island.
Why do cooperative breeders leave the group ter-
ritory?--The American Crow is the only species
known to breed cooperatively on a group ter-
ritory but leave the territory at night to attend
a distant communal roost. The typical pattern
among other cooperative breeders is for all
group members to remain on the territory both
day and night. Some species leave the territory
as a group to forage elsewhere for short inter-
vals (e.g. Green Woodhoopoes [Phoeniculus pur-
pureus], Ligon and Ligon 1990). However, this
is quite unlike our crows that leave territories
individually and show no group-related behav-
ior at distant foraging areas or at communal
roosts.
Year-round residence on breeding territories
confers at least two advantages. First, it in-
creases breeding opportunities by maintaining
ties to a successful breeding group on an es-
tablished territory. Individuals can maintain
their social position within the group, accruing
all of the benefits associated with cooperative
behavior. Related to this are benefits gained by
avoiding or delaying the need to establish dom-
inance rank when contact is made with unfa-
miliar individuals or groups away from the ter-
ritory. Young crows may be at a disadvantage
compared with older, more experienced indi-
viduals in establishing their social rank (Car-
frey 1992). The second advantage is that year-
round residence in familiar surroundings like-
ly is safer than moving to new locations during
the nonbreeding season. Intimacy with the lo-
cal landscape allows individuals to become ac-
quainted with dangers (e.g. predators) and to
become familiar with the location of safe ha-
vens and local food sources. The single factor
that limits year-round residence on territories
most often appears to be the consistent avail-
ability of adequate food supplies.
When food supplies on a territory fail to
meet energetic or nutritional requirements, the
territory holder(s) must leave to forage else-
where. A common solution to the problem is
seasonal abandonment of the territory, but this
carries a serious disadvantage in that the ter-
ritory must be reclaimed at the onset of each
breeding season. This costly and time-consum-
ing process holds the serious potential for ter-
ritory loss. Short-term diurnal abandonment of
the territory is another approach to deal with
seasonal variation in food availability. This ap-
proach is used by other cooperatively breeding,
territorial species. During the nonbreeding
season, groups of Green Woodhoopoes under-
take brief forays to foraging sites outside their
normal territory boundaries, presumably be-
cause foraging opportunities are better at these
distant sites (Ligon and Ligon 1990).
We propose that travel to distant communal
roosts by crows is another form of territory
abandonment based on the need to supplement
the foods available on the territory. In this case,
individual crows depart their territories in the
afternoon, leaving sufficient time before dark to
fulfill their foraging requirements at distant
sites. Depending on the time of year, the terri-
tory might remain occupied by some group
members, or it might be totally abandoned for
the night. In either case, it is unlikely that the
territory would be occupied by a nonresident
group. Crows leave their territories with little
or no tendency to maintain group-based be-
havior Thus, territorial contests among groups
would be unlikely late in the day when most
crows leave their territories individually to for-
age at distant sites.
Caccamise and Morrison (1986, 1988)
showed that when European Starlings travel
long distances to forage at high-quality sites
away from their DACs, they can feed there
twice for the costs of a single round trip. They
636 CACCAMISE ET AL. [Auk, Vol. 114
do this by roosting overnight near the foraging
site. Starlings use large communal roosts when
food supplies are lowest on DACs (Caccamise
1991). It is more efficient to roost overnight
than to return to the DAC whenever the for-
aging site is closer to the nighttime roost than
to the DAC. Similarly, crows can accomplish
the same efficiency when they travel to distant
foraging sites and roost nearby. Thus, we pro-
pose that large communal roosts of crows, such
as the one at Staten Island, are passive aggre-
gations that develop when many individual
crows assemble for the night at or near the
same foraging areas.
Similar patterns of movement between DACs
and distant roosts are known for at least five
communally roosting species (European Star-
lings, Morrison and Caccamise 1985; Common
Grackles, Caccamise et al. 1983; Red-winged
Blackbirds, Caccamise 1990; American Robins,
pers. obs.). Among these species, the American
Crow is the only one that breeds cooperatively.
All of the intensively studied DAC-based com-
munally roosting species, including crows, dis-
play high levels of fidelity to their DACs. Eu-
ropean Starlings use a series of foraging sites
near the roost (Caccamise and Morrison 1988)
but occasionally change roosts during a season.
In contrast, our crows always used the same
distant roost site. This difference can be ex-
plained by differences in stability of the for-
aging sites used by the two species. The sites
used by crows (landfills) probably showed lit*
tle seasonal change in quality, thus obviating
the need for crows to change sites. On the other
hand, starlings used sites that were naturally
ephemeral (e.g. fruiting trees, corn stubble, ir-
rigated turf), requiring changes in foraging
sites and the associated nearby roosts when the
quality of distant foraging sites deteriorated.
Roosting behavior in American Crows is
similar to that in the other communally roost-
ing species that display DAC-based roosting
behavior Cooperative breeding appears to
have little direct effect on the nature of com-
munal roosting in crows. Persistent fidelity to
the same DAC is interrupted by an apparent re-
source-driven need to forage at distant sites.
Thus, interpretations of the functional signifi-
cance of communal roosting that have devel-
oped from studies of other species should ap-
ply equally well to cooperative breeding in
American Crows.
ACKNOWLEDGMENTS
We thank G. Gates, R. Klimcsak, K. MacFadzean,
D. Mears, M. Robison, A. Shulman, and especially M.
Van Clef for their invaluable help in the field. We
have benefited greatly from advice and editorial
comments from K. J. McGowan and two anonymous
reviewers; we are grateful for their efforts. This is
New Jersey Agricultural Experiment Station publi-
cation No. D-08132-08-96, supported by state funds
and the United States Hatch Act.
LITERATURE CITED
ALLEN, H. L., AND L. S. YOUNG. 1982. An annotated
bibliography of avian communal roosting.
Washington State Game Department, Bothell,
Washington.
BROWN, J. L. 1987. Helping and communal breeding
in birds. Princeton University Press, Princeton,
New Jersey.
CACCAMISE, D. E 1990. Communal Starling roosts:
Implication for management. Pages 332-338 in
Proceedings of the 14th Vertebrate Pest Confer-
ence (L. R. Davis, Ed.). University of California,
Davis.
CACCAMISE, D. E 1991. European Starling fidelity to
diurnal activity centers: Role of foraging sub-
strate. Wilson Bulletin 103:13-24.
CACCAMISE, D. 17. 1993. The "patch-sitting hypoth-
esis": A parsimonious view of communal roost-
ing behavior. Wilson Bulletin 105:372-377.
CACCAMISE, D. E L. A. LYON, AND J. FISCHL. 1983.
Seasonal patterns in roosting flocks of Starlings
and Common Grackles. Condor 85:474-481.
CACCAMISE, D. F., AND D. W. MORRISON. 1986. Avian
communal roosting: Implications of "diurnal ac-
tivity centers." American Naturalist 128:191-
198.
CACCAMISE, D. E, AND D. W. MORRISON. 1988. Avian
communal roosting: A test of the patch-sitting
hypothesis. Condor 90:453-458.
CAFFREY, C. 1992. Female-biased delayed dispersal
and helping in American Crows. Auk 109: 609-
619.
CHAMBERLAIN-AUGER, J. A., P. AUGUR, AND E. G.
STRAUSS. 1990. Breeding biology of American
Crows. Wilson Bulletin 102:615-622.
GOODWIN, D. 1976. Crows of the world. British Mu-
seum (Natural History), London.
JOHNSON, R. J. 1979. Foraging distribution, habitat
relationships, and bioenergetics of roosting and
flocking Red-winged Blackbirds in central New
York. Ph.D. thesis, Cornell University, Ithaca,
New York.
KILHAM, L. 1989. The American Crow and the Com-
mon Raven. Texas A & M University Press, Col-
lege Station.
LIGON, J. D, AND S. H. LIGON. 1990. Green Wood-
October 1997] Roosting and Territoriality in Crows 637
hoopoes: Life history traits and sociality. Pages
33-65 in Cooperative breeding in birds: Long-
term studies of ecology and behavior (P. B. Sta-
cey and W. D. Koenig, Eds.). Cambridge Univer-
sity Press, Cambridge, United Kingdom.
LOMBARDO, M. iv., L. C. ROMAGNANO, i v. C. STOUFFER,
g. S. HOFFENBERG, AND H. W. IVOWER. 1989. The
use of nest boxes as night roosts during the non-
breeding season by European Starlings in New
Jersey. Condor 91:744-747.
MORRISON, D. W., AND D. •. CACCAMISE. 1985.
Ephemeral roosts and stable patches? A radio-
telemetry study of communally roosting Star-
lings. Auk 102:793-804.
MORRISON, D. W., AND D. F. CACCAMISE. 1990. Com-
parison of roost use by three species of com-
munal roostmates. Condor 92:405-412.
IVYLE, P. D., S. N. G. HOWELL, R. P. YUNICK, AND D. E
DESANTE. 1987. Identification guide to North
American passerines. Slate Creek Press, Bolinas,
California.
ROMANOWSKI, J., L. REED, D. E CACCAMISE, AND P.
STOUFFER. 1993. Colour marking American
Crows (Corvus brachyrhynchos) for individual
recognition. Notatki Ornitologiczne 34:157-161.
STOUFFER, P. C., AND D. •. CACCAMISE. 1991a. Roost-
ing and diurnal movements of radio-tagged
American Crows. Wilson Bulletin 103:387-400.
STOUFFER, P. C., AND D. E CACCAMISE. 1991b. Cap-
turing American Crows using alpha-chloralose.
Journal of Field Ornithology 62:450-453.
Associate Editor: ]. Ekman
REVIEWERS FOR THE AUK, 1997
(Continued from page 592)
L. Scott Johnson, Pierre Jouventin, Gary Kaiser,
James R. Karr, Marek Keller, William Kendall, Robert
S. Kennedy, Robert W. Kenward, Daniel M. Keppie,
Richard H. Kerbes, Paul Kerlinger, Lloyd E Kiff*,
Delbert L. Kilgore, Marcel Klaassen, Nedra K. Klein,
Richard W. Knapton, Melinda G. Knutson, Walter D.
Koenig, Erkki Korpim•iki, Burt P. Kotler, Carey Kra-
jewski, David G. Krementz, Claus K. K•nig, Melinda
S. LaBranche, Tom A. Langen, Steven Latta, Daphna
Lavee, Yvon Le Maho, James O. Leafloor, Paul L. Le-
berg, Martha L. Leonard*, Douglas J. Levey 4, Ro-
land Libois, Jan T. Lifjeld, Steve Lima, Catherine Lin-
dell, Anna K. Lindholm, fkke Lindstr•m, William
Link, Bradley C. Livezey*, Bette A. Loiselle, John
Longmire, Peter E. Lowther*, George A. Lozano,
James E Lynch, Bruce Lyon, Sheila A. Mahoney*,
Irene A. Manley*, R. William Mannan, Jeffrey S.
Marks*, Peter P. Marra, Joseph T. Marshall, Carl D.
Marti*, Larry D. Martin, Paul R. Martin, Carlos Mar-
tinez del Rio, John M. Marzluff*, Nancy E. Mathews,
Erik Matthysen, David M. Mauser, Brian McCaffery,
D. Archibald McCallum*, B. Riley McClelland, Kevin
J. McGowan, Peter K. McGregor, Frank McKinney,
M. Robert McLandress, Brian K. McNab, Scott R.
McWilliams, Neil B. Metcalfe*, Ken Meyer, Donald B.
Miles, David B. Miller, Michael R. Miller, Brian Mill-
sap*, Douglas W. Mock, Anders P. Moller, William A.
Montevecchi, Frank R. Moore, William S. Moore, Eu-
lalia Moreno, Juan Moreno, Sara R. Morris, Michael
L. Morrison, R. I. G. Morrison, Eugene S. Morton,
John M. Morton*, Helmut Mueller, Mary E. Murphy,
Michael T. Murphy, Mikko M6nkk6nen, Ruedi Na-
ger, Marc Naguib, Kenneth A. Nagy, Juan J. Negro,
Douglas A. Nelson, David N. Nettleship, Ian New-
ton*, James D. Nichols*, Jan-•ke Nilsson, Erica Nol,
Christopher Normerit, Ken Norris, Stephen Nowicki,
Thomas Nudds, Gary B. Nunn, Olof Olsson, Markku
Orell, Lewis W. Oring, Catherine Ortega, Kenneth
Otter, Kenneth C. Parkes, Cynthia S. Parr, Jeffrey D.
Parrish, Adrian Paterson, Stuart L. Paulus*, Robert B.
Payne*, Shawn D. Pearcy, Irene M. Pepperberg, Mar-
garet R. Petersen, Townsend Peterson*, Daniel Petit*,
John Piatt, Jaroslav Picman, Raymond J. Pierotti,
Theunis Piersma*, Pamela J. Pietz, Walter H. Piper,
Frank Pitelka, David Pitts, Alan Poole, Joseph P. Pos-
ton, Donald R. Powers*, Hannu Poysa, Trevor Price,
Stephen Pruett-Jones, Richard O. Prum, Kathryn L.
Purcell, Tomas Part, Robert Raikow, Ettore Randi,
Martin G. Raphael, Pamela C. Rasmussen, Claudia
Rauter, J. Michael Reed, Kerry P. Reese*, Leonard R.
Reitsma*, Thomas E. Remington, J. Van Remsen*,
Wallace B. Rendell*, Marco Restani, Eric A. Rexstad*,
Richard T. Reynolds, Judith M. Rhymer*, Heinz Rich-
her, Robert E. Ricklefs, James D. Rising*, Gary K.
Ritchison, Leslie Robb, Chandler S. Robbins, Gilbert
Roberts, Gregory J. Robertson, Raleigh J. Robertson,
Doug Robinson, Scott K. Robinson, Daniel D. Roby*,
Nicholas L. Rodenhouse*, Frank Rohwer, James J.
Roper, Kenneth V. Rosenberg, Robert N. Rosenfield*,
Jay J. Rotella, John T. Rotenberry*, Roland R. Roth,
Stephen I. Rothstein, Ian C.R. Rowley, Daniel Ru-
dolph, Carl Safina, Jorge E. Saliva*, Rex Sallabanks,
Bradley E. Sample, Michael D. Samuel, Brett Sandercock,
Ronald J. Sarno, John R. Sauer, Fred C. Schaffner,
William A. Schew, James Schieck, Ken Schmidt, Gary
D. Schnell, Hubert Schwabl, David M. Scott, Spencer
G. Sealy, William A. Searcy, Terry Shaffer, Ruth Shea,
Susan Sheaffer, Andrew L. Sheldon, Ben Sheldon,
Sara J. Shettleworth, Gerald E Shields, Kathryn E.
Sieving, Theodore R. Simons, Susan K. Skagen,
(Continued on page 687)