38
Wilson Bull., 115(1), 2003, pp. 38?44
DISTRIBUTION, ABUNDANCE, AND HABITAT AFFINITIES OF
THE COASTAL PLAIN SWAMP SPARROW
JON BEADELL,1,4 RUSSELL GREENBERG,2,5 SAM DROEGE,1 AND
J. ANDREW ROYLE3
ABSTRACT.?We examined the distribution and abundance of the Coastal Plain Swamp Sparrow (Melospiza
georgiana nigrescens) at previously occupied sites and points within potential habitat. We found Swamp Spar-
rows throughout their formerly documented range except in southern Chesapeake Bay. Swamp Sparrows were
most common in the Mullica River region of New Jersey where we detected individuals at 78% of systematically
chosen points with a mean count of 4.1 birds/point. The percentages of points with positive detections in the
regions of Delaware River (39%), eastern Delaware Bay (23%), western Delaware Bay (34%), and Tuckahoe
River (31%) were lower. The mean count of birds/point was between 0.4 and 0.6 in these regions. A higher
resolution Poisson model of relative abundance suggested that the greatest concentrations of Swamp Sparrows
occurred not only in the Mullica River area but also along northwestern Delaware Bay. Regression analysis of
Swamp Sparrow counts and habitat features identified shrubs (Iva frutescens and Baccharis halimifolia) as a
key habitat component. By applying density estimates generated by DISTANCE (Thomas et al. 1998) to the
approximate area of potential shrub habitat along Delaware Bay, we estimated that the core population of Coastal
Plain Swamp Sparrows was less than 28,000 pairs. We recommend that the Coastal Plain Swamp Sparrow be
listed as a subspecies of concern by state and local governments because of its relatively small population size,
restricted distribution in the mid-Atlantic region, and narrow habitat requirements. Received 23 April 2002,
accepted 13 November 2002.
The Coastal Plain Swamp Sparrow (Melos-
piza georgiana nigrescens) is one of three
subspecies of Swamp Sparrow, but unlike the
more common races, M. g. nigrescens is re-
stricted to a small region of the Mid-Atlantic
coastal bay ecosystem. Originally described
from specimens collected in the Nanticoke
River marshes during 1947 (Bond and Stewart
1951), M. g. nigrescens occurs only in estu-
arine marshes of the upper Chesapeake and
Delaware bays and in large river drainages
along the Atlantic coast of New Jersey north
to the tidal portions of the Hudson River
(Greenberg and Droege 1990). Subsequent
analysis has shown that M. g. nigrescens is
morphologically distinct from all other
Swamp Sparrows (Greenberg and Droege
1990). The nigrescens race is one of the few
vertebrate subspecies endemic to the mid-At-
1 U.S. Geological Survey, Patuxent Wildlife Re-
search Center, 12100 Beech Forest Rd., Laurel, MD
20708-4038, USA.
2 Migratory Bird Center, National Zoological Park,
Washington, DC 20008, USA.
3 U.S. Fish and Wildlife Service, Div. of Migratory
Bird Management, Laurel, MD 20708, USA.
4 Current address: Molecular Genetics Lab., Smith-
sonian Inst., 3001 Connecticut Ave. NW, Washington,
DC 20008, USA.
5 Corresponding author; e-mail: antbird@erols.com
lantic region, but investigations into its biol-
ogy and status are lacking.
Within mid-Atlantic estuaries, Swamp
Sparrows (M. g. nigrescens unless otherwise
noted) are restricted to an often narrow band
of shrubby habitat at the interface of upland
and high marsh. This habitat is particularly
susceptible to recent changes such as rising
sea level, ditching, impoundment, the invasion
of the reed Phragmites australis, and the de-
velopment of roads, farm fields, and housing
tracts. Studies conducted during the late 1980s
identified breeding populations of up to 70
singing males at several sites in the upper
Chesapeake Bay (RG unpubl. data). Subse-
quent visits to these and other sites with for-
merly documented breeding populations re-
vealed marked reductions of Swamp Sparrow
numbers in Maryland. Given significant de-
clines of similar marsh-nesting subspecies, in-
cluding the Cape Sable Seaside-Sparrow (Am-
modramus mirabilis; Pimm et al. 1996),
Dusky Seaside-Sparrow (A. nigrescens; Sykes
1980), and San Francisco Bay Song Sparrow
(Melospiza melodia spp.; Marshall and De-
drick 1994), we undertook a standardized sur-
vey of the Chesapeake and Delaware bays to
examine the current status of Coastal Plain
Swamp Sparrows. We surveyed previously
occupied sites and sites within potential
39Beadell et al. x COASTAL PLAIN SWAMP SPARROW
breeding habitat to evaluate range changes,
characterize habitat associations, identify re-
gions of potential management importance,
and generate baseline population estimates.
METHODS
Range prior to 2000.?We obtained records
from the Breeding Bird Survey (1966?1999;
Patuxent Wildlife Research Center, PWRC),
state Breeding Bird Atlases (Delaware, 1983?
1987; Maryland, 1983?1987; New Jersey,
1993?1997), published reports (Harlow 1907,
Stone 1937, Stewart and Robbins 1958,
Moore 1989, West 1993, Clapp 1997, Mc-
Cann and Battin 1999, Hess et al. 2000),
Maryland miniroute data, migration cards,
nest records, stomach content cards (data at
PWRC), specimen records (Univ. of Califor-
nia, Berkeley), and amateur birders through-
out the study area (see acknowledgments). We
considered only records from the coastal plain
between 15 May and 31 August in order to
avoid taxonomic confusion between M. g. ni-
grescens and migrants of other Swamp Spar-
row races (Mowbray 1997). We surveyed sites
previously occupied by Swamp Sparrows in
order to characterize their current range more
completely, but we did not include these data
in regional comparisons.
Survey design.?We surveyed six distinct
physiogeographic regions (Chesapeake Bay,
Delaware River, western Delaware Bay, east-
ern Delaware Bay, Tuckahoe River, and Mul-
lica River). We excluded regions in northern
New Jersey and the Piedmont west of Ches-
apeake Bay because morphological characters
of birds in these areas fall between those of
M. g. nigrescens and M. g. georgiana (Green-
berg and Droege 1990; RG unpubl. data). We
also excluded marshes within Blackwater Na-
tional Wildlife Refuge (NWR) and the re-
mainder of southeastern Chesapeake Bay be-
cause at the time of this study few records
existed for this region and the vast marshes
would have been over represented in our sam-
pling.
We identified potential marsh habitat along
Chesapeake Bay and its major tributaries us-
ing USGS 1:24,000 topographical maps. We
randomly allocated survey effort to these
marsh sites proportionate to their area. To en-
sure adequate coverage of northern Chesa-
peake Bay, which has numerous and contin-
uous records of Swamp Sparrows, we con-
strained our sampling by surveying 60 marsh
sites north of Annapolis and 60 to the south.
Of these, only 70 were accessible. We fol-
lowed a similar method to identify 18 marsh
sites along the Delaware River from Wilming-
ton to Philadelphia.
Swamp Sparrow habitat use is poorly doc-
umented. When possible, surveyors located a
point within shrubby brackish habitat (Green-
berg and Droege 1990) at each marsh site in
order to maximize the chance of detecting a
Swamp Sparrow. In cases where sites were in
close proximity, surveyors located points
$400 m apart to prevent counting birds twice.
Along the shores of Delaware Bay, coastal
marsh forms a continuous band of potential
habitat. Therefore, in this area, we systemati-
cally assigned points at 400-m intervals along
all roads traversing marsh habitat. Droege
(1990) discussed the validity of extrapolating
data from roadside surveys, but in our expe-
rience, roads in coastal marsh were similar to
dikes or upland barriers, which were common
in the managed wetlands surveyed. Along the
western shore of Delaware Bay, we surveyed
all accessible roadside points (n 5 176) be-
tween Lewes and Delaware City. Along east-
ern Delaware Bay, we surveyed every other
point from the Delaware Memorial Bridge
east to the confluence of the Mullica and Wad-
ing rivers. Because of the geographical isola-
tion of points in the Tuckahoe River (n 5 16)
and Mullica River (n 5 9) areas, we consid-
ered these regions to be distinct from the east-
ern shore of Delaware Bay (n 5 106 points).
We did not survey the extensive marshes
along the Atlantic Coast because prior records
do not exist for this region and spot checking
revealed no evidence of Swamp Sparrows.
Point counts.?We performed a single 10-
min count at each point between 05:00 and
10:00 EST during periods of low wind (,29
kph) and no rain. We preferred wider sam-
pling to replication since this is a more effi-
cient means of reducing total variability in in-
dex surveys (Link et al. 1994). We conducted
the survey between 25 May and 7 July 2000,
the approximate period of peak singing (RG
unpubl. data). Although singing decreased as
pairs began nesting, behaviors associated with
nesting (e.g., chipping and mobbing) helped
maintain detectability. We mapped all Swamp
40 THE WILSON BULLETIN x Vol. 115, No. 1, March 2003
Sparrows detected within a 100-m radius, re-
corded substrate use when detected visually,
and determined sex based on crown pattern
and behavioral cues (Greenberg 1988). We
used detections of Swamp Sparrows outside
of the circle and outside of standard count pe-
riods in order to establish a more complete
geographical range, but we excluded these ob-
servations from density estimates and com-
parisons of relative abundance. At each sur-
vey point, we estimated the percent area with-
in a 100-m radius circle described by each of
eight habitat categories: nonmarsh, open wa-
ter, mudflat-beach, shrub, Phragmites, rush-
reed-grass .0.5 m, rush-reed-grass ,0.5 m,
and other. Nonmarsh included forest, paved or
gravel surfaces of roads, and farm fields. Oth-
er included broad-leaved emergent plants such
as arrow arum (Peltandra virginica), picker-
elweed (Pontederia cordata), and spatterdock
(Nuphar luteum). We distinguished Phragmi-
tes from other reeds because it is structurally
distinct (taller) and it is an invasive species
whose presence has been implicated in the de-
cline of other marsh species (Benoit and As-
kins 1999). Distinguishing between reeds and
grasses of different heights served to grossly
differentiate marshes dominated by short spe-
cies such as salt-meadow grass (Spartina pat-
ens) and the short form of smooth cordgrass
(S. alterniflora) from marshes dominated by
taller species such as the tall form of smooth
cordgrass.
Density.?We estimated density of Swamp
Sparrows within the continuous band of hab-
itat along Delaware Bay using DISTANCE
(Thomas et al. 1998). Variance in counts be-
tween points in other regions was too high to
make meaningful predictions. DISTANCE
models detection probability as a function of
distance between bird and observer, thereby
accounting for potential decreases in detection
efficiency with increasing distance from the
observer. This approach assumes that detec-
tion is perfect within the smallest distance in-
terval. We grouped observations into 20-m in-
tervals to reduce the effects of error associated
with observers? distance estimates. We used
only detections of adult male birds within a
100-m circular plot to generate models for the
eastern and western shores of Delaware Bay.
The best model, selected using Akaike?s In-
formation Criterion (Akaike 1973), incorpo-
rated a uniform key function plus a cosine se-
ries expansion (Buckland et al. 1993).
Statistical methods.?We used SYSTAT 7.0
(SPSS 1997) to perform tests on data from
randomly and systematically chosen points.
Because of the nonnormal distribution of the
data, we used the nonparametric Kruskal-Wal-
lis test to judge regional differences in mean
counts of birds. Differences were deemed sig-
nificant if P , 0.05. We modeled the relation-
ship between Swamp Sparrow counts and
habitat variables using Poisson regression
(Vincent and Haworth 1983), which is a spe-
cial case of the General Linearized Model
(McCullagh and Nelder 1989). We removed
the covariate ??other?? from the model because
habitat variables are compositional (i.e., the
sum of fractional areas devoted to each habitat
character is equal to one).
We created a map of relative abundance us-
ing a Poisson model with a spatially correlated
random effect as described by Royle et al.
(2001). Under this model, the correlation be-
tween expected count at any two sites decreas-
es as an exponential function of distance be-
tween those sites. Relative abundance on the
map represents the expected mean count that
one would observe at a given location if suit-
able habitat exists.
RESULTS
Distribution.?The current distribution of
Swamp Sparrows matches the distribution of
past breeding season records except for the
absence of birds along southern Chesapeake
Bay. Swamp Sparrows occurred along the
shores of Delaware Bay where suitable habitat
exists (Fig. 1). The population extended into
marshes of the Delaware River and the large
rivers along the southern Atlantic coast of
New Jersey but not along the Atlantic coast
of Delaware.
We detected Swamp Sparrows at 78% of
points (n 5 9) in the Mullica River drainage
and this rate of detection was significantly
higher than in any other region (Table 1; H 5
18.043, P 5 0.001). No significant differences
existed among other regions (Table 1). We did
not include the Chesapeake Bay region in this
test because some randomly chosen marsh
sites included little or no habitat even though
USGS maps indicated otherwise. Detection of
Swamp Sparrows in this region was low (7%,
41Beadell et al. x COASTAL PLAIN SWAMP SPARROW
FIG. 1. Distribution of Coastal Plain Swamp Spar-
rows observed in 2000. Closed circles represent all
records of Swamp Sparrows generated during point
counts or otherwise. Open circles represent sites (ran-
domly chosen or previously occupied) at which
Swamp Sparrows were not detected during a 10-min
point count.
FIG. 2. Relative abundance of Coastal Plain
Swamp Sparrows showing concentrated populations in
the Mullica River region and northwest Delaware Bay.
Shading corresponds to the number of males predicted
to be detected during a 10-min point count in the major
regions of estuarine emergent and shrub-scrub wet-
lands and palustrine emergent and shrub-scrub wet-
lands (National Wetlands Inventory; www.nwi.fws.
gov).
TABLE 1. Indices of abundance of Coastal Plain Swamp Sparrows based on 10-min, 100-m radius point
counts performed in 2000. Counts and detection frequencies within the Mullica River region were significantly
greater than in all other regions.
Region
All points
n
% Points
with positive
detection
Maximum
count
Mean count
(6 SE)
Points with at least
one positive detection
n
Mean count
(6 SE)
Chesapeake Bay
Delaware Bay (east)
Delaware Bay (west)
Delaware River
Mullica River
Tuckahoe River
70
106
176
18
9
16
7
23
34
39
78
31
3
5
5
3
11
2
0.1 6 0.1
0.5 6 0.1
0.6 6 0.1
0.6 6 0.2
4.1 6 1.3
0.4 6 0.2
5
24
60
7
7
5
1.6 6 0.8
2.2 6 0.5
1.7 6 0.2
1.6 6 0.6
5.3 6 1.6
1.4 6 0.6
0.1 birds/point, n 5 70; Table 1). We located
only 8 Swamp Sparrows in Chesapeake Bay
during random surveys and we found only 4
by any means south of Annapolis.
Given the lack of specific habitat preference
information for Swamp Sparrows, surveyors
may not have conducted point counts in ap-
propriate locations. Therefore, a regional com-
parison of Swamp Sparrow counts at points
where at least one Swamp Sparrow was de-
tected (i.e., occupied points) may be a better
indicator of regional differences in distribu-
tion. Mean counts of birds at occupied points
along the Mullica River (5.3 6 1.6 SE) were
significantly higher than in all other regions
(H 5 15.214, P 5 0.004; Table 1). No signif-
icant differences existed among other regions.
A higher resolution map of relative abun-
dance generated by a Poisson model of pre-
dicted counts revealed that areas of greater
abundance extended farther to the mouth of
Delaware Bay along the western shore com-
pared to the eastern shore (Fig. 2). In addition,
this model indicated that regions of high abun-
dance occurred not only in the Mullica River
area but also in the northwestern portion of
Delaware Bay between Taylor?s Bridge, Del-
aware, and Salem, New Jersey.
Habitat affinity.?Counts of Swamp Spar-
rows were significantly higher at plots with a
greater proportion of shrub (Z 5 3.460, P ,
0.001; Table 2). The presence of open water
42 THE WILSON BULLETIN x Vol. 115, No. 1, March 2003
TABLE 2. Relationship between counts of Coastal
Plain Swamp Sparrows and habitat features as deter-
mined by a Poisson regression model. Swamp Spar-
rows were positively associated with shrubs, most
commonly marsh-elder and saltbush.
Habitat feature Z P
Nonmarsh
Water
Beach-mudflats
Shrub
Phragmites
Reed-rush-grass , 0.5 m tall
Reed-rush-grass . 0.5 m tall
21.388
22.080
0.244
3.460
0.981
21.202
0.220
0.165
0.038
0.807
0.001
0.327
0.230
0.826
correlated negatively (Z 5 22.080, P 5
0.038). The percentage of Phragmites in count
circles correlated positively with Swamp
Sparrow abundance, but not significantly (Ta-
ble 2). Visual detections of Swamp Sparrows
confirmed these habitat associations. Of 42
Swamp Sparrows observed directly, 33 were
found in shrubs and eight in dead or living
Phragmites.
Regional density and population esti-
mate.?We estimated the density of male
Swamp Sparrows along the eastern and west-
ern shores of Delaware Bay to be about 16/
km2 (95% CI 5 10?23/km2) and 37/km2 (95%
CI 5 27?53/km2), respectively. We generated
these density estimates from plots where the
mean percentage of unavailable habitat (roads,
agricultural fields, forest) was about 20%. Ac-
cording to National Wetlands Inventory defi-
nitions (NWI; www.nwi.fws.gov), Swamp
Sparrow habitat falls within four subclasses:
estuarine persistent emergent, estuarine broad-
leafed deciduous shrub-scrub, palustrine per-
sistent emergent, and palustrine broad-leafed
deciduous shrub-scrub. To generate a gross es-
timate of the core Swamp Sparrow population,
we summed the area of wetlands of these four
categories along the eastern and western
shores of Delaware Bay (approximately 342
km2 and 375 km2 respectively; NWI) and mul-
tiplied by the density estimates above. This
yielded a population estimate of 28,000 pairs,
assuming a strong correlation between singing
males and nesting females (Greenberg 2003).
DISCUSSION
We found the Coastal Plain Swamp Spar-
row throughout its documented range, except
along the southern Chesapeake Bay. We did
not detect Swamp Sparrows along the Poto-
mac River or the Patuxent River where birds
had been observed only sporadically through-
out the 1980s and 1990s. Similarly, we found
little evidence of Swamp Sparrows at former
breeding sites along the eastern shore of Ches-
apeake Bay, including the extensive Nanti-
coke marshes from which the type specimen
originally was described. In contrast, breeding
season records indicated continual occupancy
along northwestern Chesapeake Bay since at
least 1978. Within this region, we found small
populations of Swamp Sparrows within the
Patapsco River marshes south of Baltimore, in
the area of Black Marsh and Hart-Miller Is-
land, and in two locations on the Aberdeen
Proving Ground. We found no evidence of
large populations such as that observed by RG
(unpubl. data) at Black Marsh during the
1970s.
We found several populations of Swamp
Sparrows along the Delaware River, and a
census of the John Heinz NWR at Tinicum
revealed 22 pairs. Harlow, observing Swamp
Sparrows near Philadelphia in the early 1900s,
stated that ??this species is the most abundant
bird and nests in incredible numbers?? (1907:
122). It is unlikely that our numbers reflect
those described in Harlow?s accounts.
Swamp Sparrows appeared to be most con-
centrated along the shores of Delaware Bay.
Along the western shore, they occurred almost
continuously from Lewes to Delaware City.
The estimated density of 37 males/km2 in this
region, however, was low compared to 37?100
pairs/km2 found during 1986 at Ted Harvey
WMA (Moore 1989) and 125 territories/km2
found during 1975 at Prime Hook NWR (West
1993). A Delaware Natural Heritage study at
Woodland Beach found 189 territories/km2
during 2000 (C. Heckscher pers. comm.). This
disparity in density estimates probably is at-
tributable to patchily distributed habitat and
not fluctuating population sizes. We found
Swamp Sparrows less frequently on the east-
ern shore of Delaware Bay (15 males/km2),
especially near Port Norris where previous oc-
cupation was well documented. On the other
hand, we found high concentrations of Swamp
Sparrows in the Mullica River for which we
had no prior records of occupancy.
We observed Swamp Sparrows most com-
43Beadell et al. x COASTAL PLAIN SWAMP SPARROW
monly at the upland edges of marsh using
marsh-elder (Iva frutescens) and saltbush
(Baccharis halimifolia). We also found
Swamp Sparrows in buttonbush (Cephalan-
thus occidentalis), red maple (Acer rubrum),
and mallow (Hibiscus sp.) Dikes and roads
traversing low marsh sometimes provided
shrub habitat that may not have existed oth-
erwise. In a few cases, we found Swamp Spar-
rows using wet fields adjacent to fresh or
brackish tidal marsh. Plant species from these
fields included switch grass (Panicum virga-
tum), plumegrass (Erianthus giganteus), sedg-
es (Carex spp.), and beakrush (Rhynchospora
sp.). A positive, but insignificant, correlation
existed between Swamp Sparrows and Phrag-
mites, a reed which forms dense, monotypic
stands throughout coastal plain marshes. Male
birds often used stalks as song posts and in-
dividuals sometimes foraged at the base of
dead stalks.
Extensive, large tracts of Delaware Bay
marsh are dominated by the short form of
smooth cordgrass and salt-meadow grass.
Though not significant, Swamp Sparrow
counts correlated negatively with these short
reed-rush-grasses. Canoe surveys throughout
the low interior of marshes at Bombay Hook
NWR and Mad Horse Wildlife Management
Area corroborated this negative correlation
(SD unpubl. data). Observations in the field
suggested that the actual zone of suitable
shrub habitat available to Swamp Sparrows
typically extended only 200?400 m from the
edge of the marsh-upland transition. Thus, our
population estimate, which is derived from
habitat estimates that include low marsh,
probably is biased high.
We found only 34 Swamp Sparrows within
the Chesapeake Bay region; however, subse-
quent searching indicates that several small
populations exist in the Blackwater NWR (P.
Marra pers. comm.). The extent to which the
population penetrates the Pine Barrens (New
Jersey) or extends north toward the Hudson
River is undocumented at this time, but RG
(unpubl. data) observed very few in Hudson
Estuary tidal marshes. It currently is impos-
sible to produce a rigorous global estimate of
the Swamp Sparrow population. However, if
we double our estimate of 28,000 (for the core
Delaware Bay population) to 56,000 pairs to
account for peripheral populations in the
Chesapeake and northern New Jersey, then the
Coastal Plain Swamp Sparrow population is
on the same order of magnitude as that of the
San Francisco region Song Sparrows (Mar-
shall and Dedrick 1994), which have been
designated federal subspecies of special con-
cern.
Swamp Sparrows occupy a narrow band of
vulnerable habitat within a limited geographic
area. Coastal wetlands in the states of Mary-
land, Delaware, and New Jersey have declined
by approximately 50% since the late 1700s
(Dahl 1990) and development pressures will
result in future habitat loss. In addition, the
Environmental Protection Agency predicts a
50% chance that sea level will rise 58 cm at
Lewes, Delaware, by 2100 (Environmental
Protection Agency 1997). Rising water levels
will inundate coastal wetlands and may threat-
en currently preserved habitat. A listing of the
Coastal Plain Swamp Sparrow with other spe-
cies of concern would facilitate the develop-
ment of research and management initiatives
before the subspecies reaches threatened or
endangered status.
ACKNOWLEDGMENTS
We appreciate the dedication of the individuals who
helped perform field surveys: R. Barber-Delach, P.
Craig, D. Dawson, L. and L. Dumont, G. Hess, C.
Hetzel, M. Huffman, R. Mason, J. McCann, D. Mc-
Govern, J. Miller, D. Mitchell, A. Pupke, T. Reed, J.
Sauer, G. Scarpulla, E. Sellers, L. Sharkey, R. She-
wack, and M. Wimer. We also thank those people who
provided records: H. Armistead, D. Bystrak, J. Chur-
chill, P. Craig, L. Davidson, L. Davis, A. Dey, L. Du-
mont, J. Gruber, B. Hartman, G. Hess, M. Iliff, J. Jan-
owski, H. Laskowski, J. Lowe, J. McCann, L. Nielsen,
R. Panza, J. Reese, B. Ringler, A. Rydgren, G. Scar-
pulla, and J. Stasz. We are especially indebted to I.
Thomas for his help in generating maps and habitat
coverages in ArcInfo. M. Whitt and an anonymous ref-
eree provided many helpful comments. We also thank
the Maryland Ornithological Society, Defenders of
Wildlife, Abbott Fund of the Smithsonian, Chesapeake
Wildlife Heritage, and the Delmarva Ornithological
Society for funding.
LITERATURE CITED
AKAIKE, H. 1973. Information theory and an extension
of the maximum likelihood principle. Pp. 267?281
in International symposium on information theory,
2nd ed. (B. N. Petran and F. Csaaki, Eds.). Aka-
deemiai Kiadi, Budapest, Hungary.
BENOIT, L. K. AND R. A. ASKINS. 1999. Impact of the
spread of Phragmites on the distribution of birds
44 THE WILSON BULLETIN x Vol. 115, No. 1, March 2003
in Connecticut tidal marshes. Wetlands 19:194?
208.
BOND, G. AND R. E. STEWART. 1951. A new Swamp
Sparrow from the Maryland Coastal Plain. Wilson
Bull. 63:38?40.
BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM,
AND J. L. LAAKE. 1993. Distance sampling: esti-
mating abundance of biological populations.
Chapman and Hall, London, United Kingdom [re-
printed 1999 by RUWPA, Univ. of St. Andrews,
St. Andrews, Scotland].
CLAPP, R. B. 1997. Egg dates for Virginia birds. Vir-
ginia Avifauna, no. 6, Virginia Society of Orni-
thology, Lynchburg, Virginia.
DAHL, T. E. 1990. Wetlands losses in the United States
1780s to 1980s. USDI Fish and Wildlife Service,
Washington, DC.
DROEGE, S. 1990. The North American Breeding Bird
Survey. Pp. 1?3 in Survey designs and statistical
methods for the estimation of avian population
trends (J. R. Sauer and S. Droege, Eds.). USDI
Fish and Wildlife Service, Washington, DC.
ENVIRONMENTAL PROTECTION AGENCY. 1997. Climate
change and Delaware. Report 230-F-97-008h, Na-
tional Technical Information Service, Washington,
DC.
GREENBERG, R. 1988. Seasonal plumage dimorphism
in the Swamp Sparrow. J. Field Ornithol. 59:149?
154.
GREENBERG, R. 2003. The use of nest departure calls
for surveying Swamp Sparrows. J. Field Ornithol.
74:12?26.
GREENBERG, R. AND S. DROEGE. 1990. Adaptations to
tidal marshes in breeding populations of the
Swamp Sparrow. Condor 92:393?404.
HARLOW, R. C. 1907. Anent the Swamp Sparrow. Ool-
ogist 24:122.
HESS, E., R. WEST, M. BARNHILL, AND L. FLEMING.
2000. The birds of Delaware. Univ. of Pittsburgh
Press, Pittsburgh, Pennsylvania.
LINK, W. A., R. J. BARKER, J. R. SAUER, AND S. DROE-
GE. 1994. Within-site variability in surveys of
wildlife populations. Ecology 75:1097?1108.
MARSHALL, J. T. AND K. G. DEDRICK. 1994. Endemic
Song Sparrows and Yellowthroats of San Francis-
co Bay. Stud. Avian Biol. 15:316?327.
MCCANN, J. M. AND W. J. BATTIN, III. 1999. An in-
ventory of Neotropical migratory landbirds at the
U.S. Army Aberdeen Proving Ground, Harford
County, Maryland. Maryland Dept. of Natural Re-
sources Wildlife and Heritage Div., Wye Mills,
Maryland.
MCCULLAGH, P. AND J. A. NELDER. 1989. Generalized
linear models, 2nd ed. Chapman and Hall, Lon-
don, United Kingdom.
MOORE, E. G. 1989. Effect of Phragmites control on
use of salt marsh impoundments by breeding pas-
serines. M.Sc. thesis, Univ. of Delaware, Newark.
MOWBRAY, T. B. 1997. Swamp Sparrow (Melospiza
georgiana). No. 279 in The birds of North Amer-
ica (A. Poole and F. Gill, Eds.). Academy of Nat-
ural Sciences, Philadelphia, Pennsylvania, and the
American Ornithologists? Union, Washington,
DC.
PIMM, S., J. CURNUTT, J. LOCKWOOD, L. MANNE, A.
MAYER, M. NOTT, AND K. BALENT. 1996. Popu-
lation ecology of the Cape Sable Sparrow (Am-
modramus maritimus mirabilis): annual report,
1996. National Biological Survey/National Park
Service, Everglades National Park, Homestead,
Florida.
ROYLE, J. A., W. A. LINK, AND J. R. SAUER. 2001.
Statistical mapping of count survey data. Pp. 625?
638 in Predicting species occurrences: issues of
scale and accuracy (J. M. Scott, P. J. Heglund, M.
Morrison, M. Raphael, J. Haufler, and B. Wall,
Eds.). Island Press, Covello, California.
SPSS, INC. 1997. SYSTAT for Windows, ver. 7.0.
SPSS, Inc., Chicago, Illinois.
STEWART, R. E. AND C. S. ROBBINS. 1958. Birds of
Maryland and the District of Columbia. U.S.
Govt. Printing Office, Washington, DC.
STONE, W. 1937. Bird studies at Old Cape May. Del-
aware Valley Ornithological Club, Philadelphia,
Pennsylvania.
SYKES, P. W. 1980. Decline and disappearance of the
Dusky Seaside Sparrow from Merritt Island, Flor-
ida. Am. Birds 34:728?737.
THOMAS, L., J. L. LAAKE, J. F. DERRY, S. T. BUCKLAND,
D. L. BORCHERS, D. R. ANDERSON, K. P. BURNHAM,
S. STRINDBERG, S. L. HEDLEY, M. L. BURT, F. F.
C. MARQUES, J. H. POLLARD, AND R. M. FEWSTER.
1998. Distance 3.5. Research Unit for Wildlife
Population Assessment, Univ. of St. Andrews, St.
Andrews, United Kingdom. Available at http://
www.ruwpa.st-and.ac.uk/distance/
VINCENT, P. J. AND J. M. HAWORTH. 1983. Poisson re-
gression models of species abundance. J. Bio-
geogr. 10:153?160.
WEST, R. L. 1993. A breeding bird census from a
Prime Hook NWR marsh. Delmarva Ornithol. 25:
11?12.