BREEDING ECOLOGY OF AMERICAN AND CARIBBEAN COOTS AT SOUTHGATE POND, ST. CROIX: USE OF WOODY VEGETATION

2006·7.3 MB·English

by  D O U G L A S B Mcnair

The Wilson Journal ofOrnithology 118(2):208—217, 2006
BREEDING ECOLOGY OF AMERICAN AND CARIBBEAN
2
COOTS AT SOUTHGATE POND, ST. CROIX:
USE OF WOODY VEGETATION
DOUGLAS B. McNAIR1’34 AND CAROL CRAMER-BURKE
—
ABSTRACT. American (Fulica americana) and Caribbean (F. caribaea) coots nested colonially atbrackish
Southgate Pond, St. Croix, United States Virgin Islands (USVI), following a 50-year rainfall event in mid-
November 2003. Breeding occurred during three time periods: seven pairs bred from 6 December to 2 January
(early), seven from 17 January to 15 February (middle), and eight from 26 April to 19 May (late) (range of
clutch initiation dates = 165 days). Hatching success was high (65.3%), but overall reproductive success was
low (27%) owing to poor brood survival. Coots built all but 2 of22 nests at the water line in sturdy crotches
of small, live white mangroves (Laguncularia racemosaf two late nests were built on remnant stubs of dead
white mangroves after water levels had sharply declined. Early pairs nested in manglars (islets ofone or more
mangroves without solid land) farther away from shore and in deeper water than middle or late pairs (65.6
versus 42.1 and 29.0 cm, respectively). Southgate Pond remains the preferred breeding site for coots on St.
Croix and the USVI. Coots have also recently nested on St. Croix at seven semi-permanentorpermanent, man-
made, freshwater ponds where they have probably been overlooked, as coots respond rapidly to changes in
water levels at semi-permanent or permanent wetlands. Predominance of non-assortative pairing at Southgate
Pond suggests that American and Caribbean coots are morphs of one species. Received 7 February 2005,
accepted 7November2005.
The Caribbean Coot (Fulica caribaea) is coastal development, Southgate Pond is still
not globally threatened (Taylor 1996), but the one of the most productive mangrove wet-
species is listed as locally endangered in the lands for birds on St. Croix (Scott and Car-
United States Virgin Islands (USVI; Indige- bonell 1986, Sladen 1992; DBM and CCB un-
nous and Endangered Species Act of 1990) publ. data). We describe coot breeding adap-
and is considered threatened throughout the tations in use of woody vegetation as nest
West Indies, especially breeding populations sites (Sugden 1979), and provide information
(Raffaele et al. 1973, 1998). Caribbean and on phenology, clutch size, and breeding suc-
American (F. americana) coots are two ofthe cess. We also present recent breeding infor-
rarest bird species that nest in wetlands ofthe mation (since 2002) on coots for seven other
USVI, including St. Croix (Beatty 1930, Raf- sites on St. Croix, formulate management
faele 1989), and their breeding ecology in the strategies (especially for Southgate Pond), and
Caribbean is poorly known (Taylor 1996, assess the taxonomic significance of pairing
Brisbin et al. 2002). In North America, Amer- between the two species.
ican Coots are associated with freshwater
marshes and low-salinity brackish wetlands METHODS
(Kantrud 1985).
During 2003-2004, we studied American
Following the largest rainfall event in over
50 years, we studied the breeding ecology of and Caribbean coots at Southgate Pond, a
17.9-ha wetland (17° 45' 29.6" N, 64° 39'
CPaornid,bbethaen laanrdgesAtmesreaiscoannalcobortasckiatshSopuotnhdgaotne 45.9" W) on St. Croix, USVI. We used the cri-
St. Croix. Although degraded by previous tteirnigauiosfhRAombeerriscoanna(ntdypBeasptAisatand(1B9)88f)rotmo dCias--
1 Div. of Fish and Wildlife, Dept, of Planning and ribbean coots (types C, D, and E) in the field.
Natural Resources, 45 Mars Hill, Frederiksted, St. A small percentage (<1.4%) ofthe males with
Croix, USVI 00840, USA. broad, high, and bulbous shields may be
2St. Croix Environmental Assoc., Arawak Bldg., white-shielded morphs of American Coots
Ste. #3, Christiansted, USVI 00820, USA. (Roberson and Baptista 1988). Types A and B
3Current address: Sapphos Environmental, Inc., 133
Martin Alley, Pasadena, CA 91105, USA. have a dark chestnut or red-brown corneous
4Corresponding author; e-mail: callus, whereas types C, D, and E lack a cal-
[email protected] lus. After becoming familiar with vocal dif-
208
McNairand Cramer-Burke • COOTS AT SOUTHGATE POND, ST. CROIX 209
ferences between the sexes (Gullion 1950), we multiplying the number of active nests by
also identified the genders of some coots at mean clutch size to derive an estimate of the
their nests. Males were larger than females total number ofeggs laid, and (2) dividing the
and had larger shields and bills, regardless of numberoffully grown and independent young
species, which agrees with expectations based (not broodsperse) by eggs laid. Fledging suc-
on size and hormonal differences between the cess (number of young fledged/number of
sexes (Gullion 1951; Fredrickson 1968, 1970). eggs hatched) was determined by dividing re-
We visited Southgate Pond twice a week productive success by hatching success. The
afterthe first nest was discovered in early Jan- number of breeding pairs was based on the
uary 2004. Nests were marked with numbered number of active nests. Coot nest density at
flagging and the location of each nest was re- Southgate Pond and the seven man-made
corded with a Global Positioning System freshwater ponds was calculated based on
(GPS) unit and plotted on a map using pond area and the number of nests or pairs
ArcView 3.2. We recorded the coot species simultaneously active at each pond. Assess-
associated with each nest and coot behavior ment of intraspecific brood parasitism (“nest-
during each nest visit. Some individuals were dumping”) followed the criteria of Post and
not identified to species because of their elu- Seals (2000).
sive behavior. Dates of clutch initiation for We recorded the following parameters at
nests found during laying were calculated by each active nest and nest site: nest height from
backdating and assuming that one egg was the water line to the top of the nest rim (cm),
laid per day (Gorenzel et al. 1982, Brisbin et length and width of outer nest cup (cm),
al. 2002). Assuming a 23-day incubation pe- length and width of inner lining (cm), water
riod (Brisbin et al. 2002), initiation dates for depth below the nest (cm), above-waterheight
nests found after laying were estimated based (cm) and greatest breadth (m) of the white
on hatch dates minus 1 day (the day on which mangrove, distance to nearest white mangrove
the first egg hatched). For failed nests, we ad- (m), distance to nearest shoreline (m), distance
justed hatch date for incomplete or under-re- to nearest active nest (m), and distance to
corded clutch sizes based on the mean clutch nearest active or inactive nest (m). For each
size and backdating from the midpoint be- pair of coots, four variables (water depth be-
tween the first and last egg dates. Because our low the nest, distance from the water line to
potential renest intervals were long, renests the top of the nest rim, height of white man-
were not assigned to any one pair of coots grove above water, distance to nearest shore-
(based on criteria in Arnold 1993). line) were adjusted to the date of clutch ini-
We used the method of Mayfield (1975), as tiation. We also noted whether white man-
modified by Johnson (1979), to calculate groves that contained nests were isolated
hatching success (based on a 23-day incuba- manglars (islets of one or more mangroves
tion period). To determine reproductive suc- without solid land) or formed a line of con-
cess, we followed the fate of individuals and nected manglars away from the shoreline. We
broods until they were fully grown and inde- used a bathymetric map of Southgate Pond to
pendent (60-70 days; Taylor 1996). Young adjust distances between nests and the shore-
coots leave the nest on the day of hatching line by taking the mean value offour distance
and broods are difficult to count accurately measurements from the —15.25 to 30.5 cm
when young birds hide in emergent vegetation contour (—0.5 to 1 foot) centered on the main
(Gullion 1956); however, emergent vegetation breeding area. We then used sine/cosine func-
was scarce at Southgate Pond. As young ac- tions to calculate an angle of 0.026 degrees,
quired juvenal plumage (~3 weeks old) they which translated to a 1.9-m change in shore-
left the breeding area for deeper water along line distance per cm drop (or rise) in water
the northwestern shore of Southgate Pond, levels. Baseline water level data (in cm) were
where different broods coalesced into larger recorded in situ from several 2-m sticks
flocks and were easier to see and count. All placed in the lowest bed of the flat-bottomed
nesting attempts had known outcomes and we pond. The water level decline was nearly con-
calculated reproductive success (number of stant throughout the study period (mean of
young fledged/number of eggs laid) by (1) 0.58 cm/day), except for one heavy rainfall
210 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 118, No. 2, June 2006
event when water levels rose 17 cm from 16 distribution, including homogeneity of vari-
to 18 April. We obtained monthly measure- ances and distribution of residuals. For all
ments of salinity at Southgate Pond using a tests, we used an a value of 0.05. Means are
temperature-compensated refractometer (ac- reported ± SD.
curate to within ± 1 ppt). From December RESULTS
2003 through July 2004, salinity increased
from 4 to 32 ppt as water levels dropped. We located 22 active coot nests at South-
To systematically sample coot breeding gate Pond during winter and spring of 2003-
habitat, we established a grid of 56 line tran- 2004. Dates ofclutch initiation ranged from 6
sects, spaced 8 m apart along north-south car- December to 19 May (165 days), with breed-
m
dinal directions from 15 west of the south- ing occurring during three periods: early (6
western shoreline and extending to the eastern December to 2 January; 27 days), middle (17
point of Southgate Pondjust beyond the main January to 15 February; 29 days), and late (26
coot colony. We randomly selected sample April to 19 May; 23 days). One nesting at-
points (n = 436) every 8 m along each tran- tempt during the late period was overlooked
sect. The last point along each transect was a (see below). We identified one pair of Carib-
point in open water beyond the vegetation far- bean Coots and five Caribbean X American
thest from shore. Waterdepth (cm), vegetation coot pairs (hereafter mixed pairs) during the
present or absent (open water), and species early period, two pairs of Caribbean Coots
composition (ifvegetation present) were sam- and four mixed pairs during the middle period
pled at each point. We used a random number (two male American and two female Carib-
generator to assign numbers 1 to 22 (i.e., cor- bean coots were sexed in two of these four
responding to the number of coot nests we mixed pairs), and three pairs of Caribbean
found) to sample points ofwater depth. Water Coots and two mixed pairs during the late pe-
depth at each sampled point that contained riod (both males were American and both fe-
vegetation was then adjusted to reflect water males were Caribbean coots). One of the
depth at the observed or estimated date of American Coots of one mixed pair during
clutch initiation for each coot nest represented each of the first two periods was type B (in-
by each random number (e.g., random number termediate, sensu Roberson and Baptista
one represents coot nest one, which initiated 1988). The other American Coots appeared to
incubation on 6 December). This procedure be type A birds. Of the three coots whose
removed the effects of declining water levels mates were not identified, two were Caribbean
so that vegetation data would be comparable and one was American.
to nest data. Clutch size decreased as the nesting season
tweTeonavsesgeestsatdiiofnfearnendceospeinn wwaatteerr adnedptahsmobneg- eprroaggreedss6e.d88(r±s =1.4—10.e5g6g,sP(r=ang0e.0=25)5-a9n,dnav—-
species of plants, we used two-tailed f-tests 16). Seventeen of 22 nests (77%) hatched at
ANOVA
and one-way (Zar 1999, StatSoft least one chick, and only 5 of the 130 eggs
2002). Because the sample sizes for five of (3.8%) that remained unbroken in the nest
the eight vegetative species/types recorded bowl throughout the normal incubation period
were small (total n = 15), we did not include failed to hatch. Daily nest survival (5) was
them in the ANOVA. We used simple linear 0.982 ± 0.008 SE and hatching success was
regression to assess the relationship between 65.3% (Mayfield method). Hatching success
water depth below nests and the date ofclutch was not related to clutch initiation dates (U =
initiation. We used a Mann-Whitney U-test to 26, Z = 1.29, P = 0.20) or any other pheno-
assess whether water depths at coot nests dif- logical or habitat variable, although successful
fered from random and to examine whether nests generally began earlier and had larger
phenological or habitat variables were related clutches, greater water depths, and were far-
to hatching outcome (success/fail). We used ther away from shore than failed nests. Forty-
nonparametric tests (Mann-Whitney U Krus- one young became fully grown and indepen-
,
kal-Wallis //, and Spearman’s rank correlation dent 60-70—days after hatching. This excludes
rs) when sample sizes were small and data did three y—oung attended by a pair ofCaribbean
not otherwise meet assumptions ofthe normal Coots that fledged from a ninth nest over-
85
McNairand Cramer-Burke • COOTS AT SOUTHGATE POND, ST. CROIX 21 1
TABLE 1. Measurements of 14 nest and nest-site parameters for 22 coot nests built in white mangroves at
Southgate Pond, St. Croix, U.S. Virgin Islands, during winter and spring of2003-2004.
Parameter Mean ± SD Range
Nest height from water line to top ofnest rim (cm) 13.5 6.9 4.8-35.5
Length ofouter nest cup (cm) 35.9 ± 8.8 25.4-61.0
Width ofouter nest cup (cm) 28.6 ± 4.2 20.3-36.2
Length ofinner lining (cm) 19.8 + 2.2 15.5-25.4
Width ofinner lining (cm) 18.0 ± 2.1 14.0-22.9
Water depth below nest (cm) 44.8 ± 17.6 15.0-78.5
Above-water height ofwhite mangrove (m)a 1.9 ± 0.6 1.0-3.7
Greatest breadth ofwhite mangrove (m)b 3.9 ± 1.1 1.5-5.
Distance to nearest white mangrove (m)c 3.2 -t- 2.4 0.0-8.
Distance to nearest shoreline (m) 48.4 -t- 26.0 10.4-98.1
Distance to nearest active nest (m) 60.4 + 59.6 18.7-308.2
Distance to nearest active nest (m)d (excluding three isolated nests) 42.7 -4- 13.5 18.7-60.2
Distance to nearest active or inactive nest (m) 42.3 ± 59.9 10.2-283.9
Distance to nearest active or inactive nest (m)d (excluding three isolated nests) 23.1 -I- 11.3 10.2-50.5
aOneoutlierexcluded(deadwhitemangrove:nest 17;height<20cm).
bTwooutliersexcluded(onedeadwhitemangroveandonelivewhitemangrove:nests 17,21;breadthnotmeasuredand =55.7m,respectively).
cOneoutlierexcluded(livewhitemangrove:nest22;distance = 127.9m).
dOneisolatednestexcludedfromeachofearly,middle,andlatenestingperiods(nests7, 13,and 17).
looked during the late period (date of clutch (forb) of Southgate Pond was widgeon grass
initiation was later than 19 May). The largest (Ruppia maritima), but this species was not
single brood observed comprised five young used as nest material. Most manglars, both
(from a mixed pair), and there were six broods white and black Avicennia germinans man-
( )
(from four mixed pairs and two Caribbean groves, were located at the east end of the
Coot pairs) with four young. Reproductive pond, where most nests were concentrated
success was 27%, and fledging success was (Fig. 1). Two rather isolated nests (7, 13) were
41.3%. near the southwestern shoreline, and the most
Nests were built along the water line in par- isolated nest (17) was near the northwestern
tially submerged, small, live white mangroves shoreline. The density ofcoot nests during the
(Laguncularia racemosa\ Table 1). Most nests three periods was 0.39-0.45 nests/ha.
were placed either in the central crotch (early Mean waterdepth at nests was 44.8 cm (Ta-
and middle periods) or in smaller crotches of ble 1) and declined throughout the breeding
outside branches (late period); two nests dur- season (early period: 65.6 cm ± 11.0; middle
ing the late period were also placed either on period: 42.1 cm ± 3.8; late period: 29.0 cm
remnants of dead white mangroves under live ± 9.5; Kruskal-Wallis H = 15.14, P < 0.001;
vegetation oron unconcealed dead white man- Fig. 2). Early nests were also farther away
groves. All nests during the early and middle from the shoreline than middle or late nests
periods had short or long ramps, while only (early period: 70.0 m ± 29.7; middle period:
two nests during the late period had ramps. 49.6 m ± 11.2; late period: 28.5 m ± 15.8;
Nests were in isolated manglars (n = 18) or Kruskal-Wallis H = 8.81, P = 0.010). Other
in rows of manglars {n = 4), but away from comparisons of nest or nest-site variables be-
mangroves that formed the outer fringes of tween early, middle, and late periods were not
Southgate Pond’s vegetated shoreline. Nests significantly different.
were located close to nest materials, the bulk Vegetation sampled at random points along
of which (excluding sticks and twigs of man- line transects composed 34.6% (n = 151) of
groves) consisted of shoreline sea purslane breeding habitat; the remainder was open wa-
(Sesuvium portulacastrum), a perennial suc- ter (n = 285), where mean water depth was
culent forb also used to construct most of the significantly greater than in vegetated areas
ramps. Seed pods ofSesbania sericea, a short- (open water: 45.8 cm ± 27.3; vegetation: 37.4
lived shrub, composed the inner nest lining of cm ± 26.2; t = 3.11, df 434, P = 0.002).
several nests. The dominant submerged plant Live white and black mangroves and dead
212 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 118, No. 2, June 2006
A
60 120 180 240 300 360 420 Meters
FIG. 1. The location of22 coot nests at Southgate Pond, St. Croix, U.S. Virgin Islands, during winter and
spring of2003-2004.
white mangroves dominated vegetation types without coot nests were similar(without nests:
within breeding habitat, and mean water 39.3 cm ± 25.7; with nests: 44.8 cm ± 17.6;
depths at dead and live white mangroves were Mann-Whitney U = 434.5, P = 0.24).
—
significantly deeper than at black mangroves Freshwater ponds. Since 2002, 1—3 pairs
(F2134 = 8.28, P < 0.001; Table 2). Mean wa- ofCaribbean Coots and mixed pairs have bred
ter depths at live white mangroves with and intermittently year-round at seven man-made,
freshwater ponds on St. Croix, which range in
size from 0.1 to 2.9 ha. The mean coot density
90:
at all sites combined for all breeding sequenc-
80:. y=61.84-0.23x,r2=0.68,P<0.001 es over the 4 years was 4.2 pairs/ha (range =
=
70:* 0.3-10.0, n 17) and apparent hatching suc-
E cess was high (16 of 24 nests based on hatch
60:
rates of the proportion of nests found). Most
breeding records occurred after the 50-year
rainfall event of mid-November 2003 filled
— the ponds. This event followed a dry period,
when a variety of plant species had colonized
20
: the bottom of many dry, or nearly dry, ponds.
101 - In addition, the Virgin Islands Agricultural
0 20 40 60 80 100 120 140 160 180
Dateofclutchinitiation(1=6December) Station Middle Pond (see McNair 2006 forlist
of pond names and their locations on St.
FIG. 2. Relationship between water depth below
Croix) was deliberately drained during winter
22 coot nests and the date ofclutch initiation at South-
gatePond, St. Croix, U.S. VirginIslands,duringwinter 2002-2003. Water levels variedbetween years
and spring of2003-2004. at several sites when nesting occurred, espe-
8
McNairand Cramer-Burke • COOTS AT SOUTHGATE POND, ST. CROIX 213
TABLE 2. Mean water depth (cm) for eight vegetation types at Southgate Pond, St. Croix, U.S. Virgin
Islands, during winter and spring of2003—2004.
Vegetationtype n Mean ± SDh
Dead Laguncularia racemosa 54 45.4 ± 26.9 A
Live Laguncularia racemosa 48 39.3 ± 25.7 A
LiveAvicennia germinans 34 23.2 ± 21.2 B
DeadAvicennia germinans 3 38.6 ± 20.6C
Sesbania sericea 4 43.9 ± 17.0C
Sesuviumportulacastrum 4 27.8 ± 28. C
Sesuviumportulacastrum on dead L. racemosa 2 52.7 ± 20.2C
Sporobolus virginicusa 2 6.8 ± 29.lc
All vegetation 151 37.4 ± 26.2
Open water 285 45.8 ± 27.3
aSeashorerushgrass.
livbeOAvveircaelnlni^a2,g1e34rm=in8a.n2s8,vePrs<us0l.i0v0e1;Largouwnscuwliatrhiadirfafecreemnotsal'e,ttePrs<(A0,.0B0)1afroersliigvneifAi.cagntelrymidnifafnesrevnter(sTuuskedye’asduLn.eqruaaclemnosHa)S.Dpost-hoctests:P=0.026for
cSamplesizetoosmalltotest.
cially at the Virgin Islands Agricultural Sta- overflow. Coots colonized the pond and began
tion Middle Pond. Live creeping burrhead laying eggs within 2-3 weeks after this sea-
(Echinodorus berteroi) was almost absent sonal wetland filled with water, typical of
there in 2004, when the pond was not used by coots after arrival on their breeding grounds
coots and emergent vegetation was restricted (Alisauskas and Arnold 1994).
to the shoreline when the water level was When conditions are suitable, Southgate
higher. Although coots nested in a variety of Pond is probably the preferred breeding site
live (five species) and dead (two species) veg- for coots on St. Croix (and in the USVI;
etation, woody (especially remnant S. sericea, McNair 2006), even though freshwater ponds,
at four ponds) vegetation ratherthan perennial each with a small number of birds, support
herbaceous vegetation was the predominant higher breeding densities (this study). Al-
nest substrate (18 of 27, 67%). Nests ranged though Southgate Pond is brackish, the num-
from 4 to 33.5 m away from the shoreline, ber ofbreeding pairs during three consecutive
and water depths below nests were generally periods from December to May did not de-
greater for nests built in woody vegetation, cline as salinity increased from low to mod-
especially S. sericea (usually 1.25-2.25 m). erately brackish; elsewhere, breeding densities
The bulky, conspicuous nests composed of typically decline as salinity increases (Kantrud
sticks of S. sericea (—90 X 65 cm) were su- 1985, Arnold 1993). Regardless, semi-per-
perficially shaped like the above-waterportion manent or seasonal wetlands are generally
of a beaver lodge. Anthropogenic disturbance preferred habitat for American Coots in North
at these seven ponds was negligible except America (Kantrud 1985, Arnold 1993, Ali-
around Carlton North Pond, where all vege- sauskas and Arnold 1994). Nests at Southgate
tation except that fringing the shoreline was Pond, which generally lacked emergent her-
cleared for a housing development in early baceous vegetation, were built in woody veg-
October 2004; however, coots continue to etation. In Saskatchewan, small, isolated, par-
breed at Carlton North Pond. tially submerged willow (Salix spp.) clumps
were used as nest sites for a substantial per-
DISCUSSION
centage (22%) ofAmerican Coot nests during
Because of a drought on St. Croix that be- a wet year (Sugden 1979), although willows
gan in 2002, the bottom of Southgate Pond were not used as nest materials. This is dif-
was dry in 2003 until water from heavy rains ferent from what we observed at Southgate
began to fill the pond in late August. None- Pond, where white mangroves served as nest
theless, the basin was only about one-quarter sites and as nest-building material; remnant or
full until a 50-year rainfall event during 10- live woody plants (especially remnant S. ser-
14 November 2003 caused Southgate Pond to icea) at freshwater ponds on St. Croix were
214 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 118, No. 2, June 2006
used similarly. In another Saskatchewan Maher 1987, Post 1990, Arnold 1993). Fur-
study, coot nests were composed of the same thermore, white mangroves offer superior
plant species that provided supportforthe nest structural support for nests (black mangroves
(cf. Sutherland and Maher 1987). lack the sturdy bowl-shaped central crotch and
Water depths below many nests in fresh- low lateral branches) and greaterconcealment.
water ponds on St. Croix were much deeper For similar reasons, American Coots in Sas-
than water depths below nests in white man- katchewan nested in live willows but not
groves at Southgate Pond. Apart from South- quaking aspens (Populus tremuloides) (Sug-
gate Pond, the most suitable freshwaterbreed- den 1979).
ing site for coots on St. Croix has been Gran- As water levels declined, nest-site selection
ard South Pond, where three pairs nested in changed; by the late period, the central crotch-
remnant Sesbania and other nest sites. Unlike es ofwhite mangroves were no longersuitable
mangroves at Southgate Pond, suitable rem- (too far above water). Nonetheless, inter-nest
nant woody vegetation at freshwater ponds distances remained similar during all three pe-
usually becomes available only when these riods, suggesting that territory sizes (which
ponds dry up and then refill with water, which were not measured) also remained similar. In-
kills the colonizing shrubs. Emergent vegeta- ter-nest distances between simultaneously ac-
tion suitable for nests at some of these ponds tive (or inactive) nests during all three periods
can be scarce, even when water levels are low. were typical of those observed for coots else-
Nests in perennial emergent forbs were float- where, although published data are unavail-
ing platforms built amongst this vegetation, able for nests limited to woody vegetation.
which is typical of coot nests in marshes Unlike what has been observed at many North
(Fredrickson 1970, Sugden 1979, Gorenzel et American sites characterized by emergent
al. 1982, Kantrud 1985, Post 1990, Alisauskas vegetation, coots at our study site built few
and Arnold 1994, Frost and Massiah 2001). non-nesting platforms (six in white man-
At Southgate Pond, water depths at coot groves), and the distribution and structure of
nests during each period were typical ofthose nesting cover at Southgate Pond did not
observed at American Coot nests on the North change over the breeding season. Given the
American mainland (Sugden 1979, Gorenzel fixed number of live white mangroves as po-
et al. 1982, Sutherland and Maher 1987, Post tential nest sites for coots at Southgate Pond,
1990, Arnold 1993), although depths during territorial behavior probably prevented any
the third period were rather shallow. Even additional coot pairs from breeding at the site.
though coots on the North American mainland The location ofcoot nests is mainly controlled
frequently nest in residual emergent vegeta- by territorial spacing, distance from shore, and
tion (Gorenzel et al. 1982, Alisauskas and Ar- the distribution and structure of nesting cover
nold 1993), in our study they probably avoid- (Gullion 1953, Sugden 1979, Sutherland and
ed using dead white mangroves as nest sites Maher 1987). Water depth, although correlat-
in deeper water at Southgate Pond because ed with distance from shore in this study, was
nests in these sites would have been exposed. probably a less important factor in nest-site
Were it not for the effects of hurricanes Hugo selection.
and Marilyn in 1989 and 1996, which killed Nest concealment in woody vegetation
many white mangroves farther from shore, must have been effective because hatching
several more pairs of coots may have used success at Southgate Pond was high. Apparent
these mangroves as nest sites. Coots also hatching success was also high at freshwater
avoided nesting in black mangroves, which sites, which is typical of American Coots
are generally located closer to shore than the (Gorenzel et al. 1982, Alisauskas and Arnold
live white mangroves they used. Water depths 1994, Brisbin et al. 2002). Intraspecific nest
at nests in white mangroves during the late parasitism was not observed at Southgate
period were similar to mean depths at black Pond or at the freshwaterponds. Fledging suc-
mangroves, suggesting that water depth at cess at Southgate Pond, although not consis-
black mangroves was otherwise acceptable to tently associated with differences in water
coots. However, coots generally prefer deeper depth, was low (<41%). This contrasts with
water farther from shore (Sutherland and apparent fledging success at freshwater sites
McNairand Cramer-Burke • COOTS AT SOUTHGATE POND, ST. CROIX 215
(this study), and that in North America, which tal Association) remains favorable habitat for
is generally high (>50%; Alisauskas and Ar- nesting coots, even though environmental
nold 1994). Most broods observed at South- degradation has diminished this brackish pond
gate Pond consisted of 2-3 birds, lower than to <50% of its original size (Gaines 2004,
the number typically observed at freshwater Gaines and Gladfelter 2004). The most diffi-
ponds (7 of 13 broods had >4 fledged young). cult task at Southgate Pond is to maintain ap-
Thus, we speculate that brood losses within 5 propriate water levels for coot nest initiation
days after hatching exceeded 50% at South- during seasons and years when rainfall is in-
gate Pond. Low survivorship of young also sufficient. We endorse Gaines and Gladfelter’s
occurred during the early brood period for (2004:54-56) two major recommendations for
White-cheeked Pintails {Anas bahamensis) at water management to prolong the aquatic
Humacao, Puerto Rico (F. J. Vilella pers. phase ofSouthgate Pond: (1) divert water into
comm.), where most losses were attributed to the pond, and (2) raise the maximum water
rats (Rattus spp.), Great Egrets (Ardea alba), depth from —103 to —138 cm. Manipulation
and Black-crowned {Nycticorax nycticorax ofwater levels should favor nesting coots and
)
and Yellow-crowned {Nyctanassa violacea) other wetland birds, although it may eliminate
night-herons. All of these potential predators species that nest in terrestrial sites. During its
were present at Southgate Pond. dry phase, two species of conservation con-
Despite low reproductive success at South- cern on St. Croix may nest at Southgate Pond:
gate Pond, the long intervals between breed- Wilson’s Plover {Charadrius wilsonia) and
ing periods and the similar number of pairs Least Tern {Sterna antillarum). However, both
breeding during each period suggest that some species breed at more than 10 sites and are
middle and late period nests were probably not as rare as coots. Furthermore, Southgate
second or third broods rather than renests. Pond is the best-documented site for coots in
Presumed success of second nesting attempts the eastern Caribbean (McNair 2006). When
also occurred at three of the seven freshwater water levels are sufficient, the brackish habitat
ponds. Nesting during the late period at at Southgate Pond may be similar to that of
Southgate Pond appeared to be possible be- brackish impoundments along the northern
cause of heavy rainfall that occurred from 16 Gulfcoast ofthe United States (e.g., an abun-
to 17 April, when water levels rose 17 cm, dance of sea purslane and widgeon grass),
allowing coots to reset their breeding clock where coots are abundant (Swiderek et al.
despite an overall drop in water level (14 cm) 1988).
since the middle breeding period. Before the At the seven man-made, freshwater ponds,
50-year rainfall event ofmid-November 2003, piped water is generally the best management
coots probably last nested at Southgate Pond option to maintain stable, generally high water
in 2001, following the previous torrential rain- levels. The most suitable freshwater site in the
fall event of8 May when water filled the pond eastern Caribbean (Barbados) is man-made
(CCB unpubl. data). This opportunistic, multi- Marshall’s Pond, which is dominated by Echi-
brooded breeding response to aquatic periods nodorus berteroi (Frost and Massiah 2001; M.
resulting from torrential vernal and autumnal D. Frost pers. comm.), the herbaceous species
rainfalls in an otherwise semi-arid environ- used most frequently for nest sites on St.
ment may allow coots to overcome generally Croix. Maintaining stable water levels at the
low reproductive success on St. Croix. Nev- best site on St. Croix (Granard South Pond),
ertheless, three breeding periods during one as well as at the other ponds, should generally
aquatic phase is probably exceptional (DBM favor E. berteroi and other species with sim-
and CCB unpubl. data). How frequently and ilar vegetative characteristics. Woody vegeta-
successfully coots breed at Southgate Pond tion would no longer compose the dominant
and freshwater sites on St. Croix in the future nest sites because stable water levels would
is currently being determined through an on- generally prevent woody plants such as S. ser-
going wetlands bird-monitoring scheme. icea from becoming established except along
—
Management recommendations. South- the immediate shorelines of these ponds.
—
gate Pond (now part of the Southgate Coastal Caribbean Coot systematics. The taxo-
Reserve owned by the St. Croix Environmen- nomic status of the Caribbean Coot requires
—
216 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 118, No. 2, June 2006
further investigation (Roberson and Baptista ofNatural Sciences ofPhiladelphia, Philadelphia,
1988; also Gullion 1951, Phillips 1967, Payne Pennsylvania.
and Master 1983, Clark 1985, Taylor 1996). Brisbin, I. L., Jr., H. D. Pratt, and T. B. Mowbray.
2002. AmericanCoot Fulicaamericana andHa-
AvapiplaerdeanttSnoount-hagsastoretPatoinvde,pwahierirnegboofthcotoytpsepsreo-f iwcaai,ianon.C6o9o7t.(F. alai).(The Birds ofNort)h Amer-
coots occurred. One-half of the pairs at fresh- Clark, C. T. 1985. Caribbean Coot? Birding 17:84-
water sites on St. Croix were paired non-as- 88.
sortatively. Furthermore, at least some mixed Fredrickson, L. H. 1968. Measurements ofcoots re-
pairs successfully raised young, especially at lated to sex and age. Journal ofWildlife Manage-
Southgate Pond, indicating that the two types ment 32:409-411.
Fredrickson, L. H. 1970. Breeding biology ofAmer-
of coots can produce viable offspring (Gill ican Coots in Iowa. Wilson Bulletin 82:445-458.
1964, Payne and Master 1983, Bond 1984). Frost, M. D. and E. B. Massiah. 2001. Caribbean
Thus, American and Caribbean coots may Coot (Fulica caribaea): the return of a former
compose one species with variant, intergraded breeding resident bird. Journal of the Barbados
phenotypes of which A and E birds represent Museum and Historical Society 47:85-90.
the extreme types. Although some birds can Gaineosl,ogAy.aGn.,dJhry.d2r0o0l4o.gyThoefSaonutahrgiadtelawnadtsecrasphee.d:SgCeR-
be individually recognized in the field, an ac- Technical Report, no. 2. The Coast & Harbor In-
curate assessment of phylogenetic relation- stitute, Woods Hole, Massachusetts.
ships and the taxonomic status of American Gaines, A. G., Jr., and E. H. Gladfelter. 2004. The
and Caribbean coots will require studies based Southgate Coastal Reserve: a strategy for man-
on morphological and genetic analyses along agement and implementation. SCR Technical Re-
with observations ofmating behavior and pair port, no. 1. The Coast & Harbor Institute, Woods
Hole, Massachusetts.
bonds of marked birds. This will also require Gill, F. B. 1964. The shield color and relationshipsof
confirming identification of shield character- certain Andean coots. Condor 66:209-211.
istics and correlating them with other mor- Gorenzel, W. R, R. A. Ryder, and C. E. Braun.
phological measurements. 1982. Reproduction andnestsitecharacteristicsof
American Coots at differentaltitudesinColorado.
ACKNOWLEDGMENTS Condor 84:59-65.
Gullion, G. W. 1950. Voicedifferencesbetweensexes
The U.S. Fish and Wildlife Service provided partial in the American Coot. Condor 52:272-273.
financial support to the Division ofFish and Wildlife, Gullion, G. W. 1951. The frontal shield ofthe Amer-
U.S. Virgin Islands (Federal Aid Program, Pittman- ican Coot. Wilson Bulletin 63:157-166.
Robertson Wetlands Project, W15). S. L. Fromer and Gullion, G. W. 1953. Territorial behavior of the
L. D. Yntema shared their bird observations; A. G. American Coot. Condor 55:169-186.
Gainescontributeddataon salinitiesatSouthgatePond Gullion, G. W. 1956. An observation concerning the
and provided a digitized bathymetric map ofthis site; validity ofcoot brood counts. Journal ofWildlife
D. B. Nelthrop and D. M. Schuster, among otherland- Management 20:465-466.
owners, provided access to their freshwater ponds; J. Johnson, D. H. 1979. Estimating nest success: the
A. Collazo and F. J. Vilella reviewed a penultimate Mayfield method and an alternative. Auk96:651-
draft ofthe manuscript, and T. W. Arnold, F. E. Hayes, 661.
andananonymous individual reviewedearlierversions Kantrud, H. A. 1985. AmericanCoothabitatinNorth
ofthe paper.
Dakota. Prairie Naturalist 17:23-32.
LITERATURE CITED Mayfield, H. F. 1975. Suggestionsforcalculatingnest
success. Wilson Bulletin 87:456-466.
Alisauskas, R. T. and T. W. Arnold. 1994. American McNair, D. B. 2006. Reviews ofthe status ofAmer-
Coot. Pages 126-143 in Migratory shore and up- ican Coot(Fulicaamericana) andCaribbeanCoot
land game bird management in North America(T. (Fulica caribaea) in the United States Virgin Is-
C. Tacha andC. E. Braun, Eds.). InternationalAs- lands. North American Birds 59:678-684.
sociation of Fish and Wildlife Agencies, Allen Payne, R. T. and L. L. Master. 1983. Breeding ofa
Press, Lawrence, Kansas. mixed pair of white-shielded and red-shielded
Arnold, T. W. 1993. Factors affecting renesting in American Coots in Michigan. Wilson Bulletin95:
American Coots. Condor 95:273-281. 467-469.
Beatty, H. A. 1930. Birds ofSt. Croix. Journal ofthe Phillips, A. R. 1967. Some Antilleancoots (Fulica)in
DepartmentofAgricultureofPuertoRico 14:135 the Cambridge University and British Museums.
150. Bulletin ofthe British OrnithologicalClub 87:35-
Bond, J. 1984. Twenty-fifth supplement to the check- 36.
list of birds of the West Indies (1956). Academy Post, W. 1990. AmericanCootsnestinSouthCarolina
McNairand Cramer-Burke • COOTS AT SOUTHGATE POND, ST. CROIX 217
aftera 35-yearinterlude, and a summary ofSouth Sladen, F. W. 1992. Abundance and distribution of
Carolina coot nidiology. Chat 54:9-11. waterbirds in two types ofwetlands on St. Croix,
Post, W. and C. A. Seals. 2000. Breeding biology of U.S. Virgin Islands. Ornitologia Caribena 3:35-
the Common Moorhen in an impounded cattail 42.
marsh. Journal ofField Ornithology 71:437-442. StatSoft, Inc. 2002. Statistica base, ver. 6.1 forWin-
Raffaele,H. 1989. AguidetothebirdsofPuertoRico dows. StatSoft, Inc., Tulsa, Oklahoma.
and the Virgin Islands, 2nd ed. Princeton Univer- Sugden, L. G. 1979. Habitat use by nesting American
sity Press, Princeton, New Jersey. Coots in Saskatchewan parklands. WilsonBulletin
Raffaele, H. A.,M.J. Velez, R. Cotte,J.J. Wehlan, 91:599-607.
E. R. Keil, and W. Cumpiano. 19—73. Rare and Sutherland, J. M. and W. J. Maher. 1987. Nest-site
endangered animals ofPuerto Rico acommittee selection ofthe American Coot in the aspen park-
report. U.S. Department ofAgriculture-Soil Con- lands ofSaskatchewan. Condor 89:804-810.
ssoeurrvcaetsi,onCSoemrmvoicneweaanldtDhepoafrPtumeernttooRficNoa.tural Re- Swiderek, P. K., A. S. Johnson, P. E. Hale, and R. L.
Joyner. 1988. Production, management, and wa-
RaffaRealfef,aeHl.e,.J.19W9i8l.eyA,gOu.idGeartroitdhoe,bAi.rdsKeoiftht,heanWedsJt. terfowluseofseapurslane.GulfCoastmuskgrass,
Indies. PrincetonUniversityPress,Princeton,New and widgeongrass in brackish impoundments.
Pages 441-457 in Waterfowl in winter (M. W.
Jersey.
Roberson, D. andL. F. Baptista. 1988. White-shield- Weller, Ed.). University ofMinnesota Press, Min-
ed coots in North America: a critical evaluation. neapolis.
American Birds 42:1241-1246. Taylor, P. B. 1996. Family Rallidae (rails, gallinules,
Scott, D. A. and M. Carbonell. 1986. A directory and coots). Pages 108-209 in Handbook of the
of Neotropical wetlands. International Union for birds ofthe world, vol. 3: Hoatzin to auks (J. del
Conservation of Nature and Natural Resources, Hoyo, A. Elliott, and J. Sargatal, Eds.). Lynx Ed-
Cambridge, United Kingdom, and International icions, Barcelona, Spain.
Waterfowl and Wetlands Research Bureau, Slim- Zar, J. H. 1999. Biostatistical analysis, 4th ed. Pren-
bridge, United Kingdom. tice-Hall, Upper Saddle River, New Jersey.