Content uploaded by Chris J Hassell
Author content
All content in this area was uploaded by Chris J Hassell on Mar 14, 2016
Content may be subject to copyright.
Introduction
Two subspecies of Gull-billed Tern (Sterna nilotica) occur in
northern Australia. Subspecies macrotarsa is a large, pale
Australian subspecies characteristic of inland wetlands
(Blakers et al. 1984; Higgins and Davies 1996) and was long
thought to be the only subspecies in Australia. In the 1970s
Johnstone (1977) recorded the first Australian specimen of a
second, smaller subspecies, which he collected at the mouth
of the Lawley River in the Kimberley and identified as
S. n. affinis. Since then there has been some uncertainty
about the status of affinis in Australia. Johnstone and Storr
(1998) considered affinis to be only a casual visitor to north-
western Australia and accepted only one record. However,
when Schodde (1991) reported two further specimens from
the Alligator River, Northern Territory, he noted that small
Gull-billed Terns of similar appearance had been seen regu-
larly in the estuaries of the Alligator Rivers between October
and April, and therefore proposed that affinis is a regular
wintering migrant to Australia.
There is also uncertainty about the taxonomic status of
migrant Asian Gull-billed Terns in Australia. Although they
have usually been treated as subspecies affinis, Higgins and
Davies (1996) noted that measurements of the three
Australian specimens available do not correspond perfectly
with published measurements of affinis from overseas. In
addition, taxonomy of the Gull-billed Terns occurring in
eastern Asia (the presumed source of the migrants found in
Australia) is unsettled, with regional guides and checklists
proposing a variety of treatments on the basis of scanty
primary literature. In theory the Asian migrants occurring in
Australia could belong to any one of the following three sub-
species: (1) Nominate nilotica, which is usually regarded as
breeding from Europe, through the Middle East and
Kazakhstan, to at least north-western China. It has been said
that its breeding range probably extends into Mongolia
(Dement’ev and Gladkov 1951; Vaurie 1965) and possibly as
far east as Manchuria (Vaurie 1965; Étchécopar and Hüe
1978). Although no newer primary literature on this problem
appears to have been published, Manchuria has since been
treated as the eastern limit of the range without qualification
(e.g. Cramp 1985). Given that some shorebirds from
Mongolia and Manchuria migrate to Australia (Higgins and
Emu, 2005, 105, 145–158
10.1071/MU04045 0158-4197/05/020145© Royal Australasian Ornithologists Union 2005
Danny I. RogersA,F, Peter CollinsB, Rosalind E. JessopC, Clive D. T. MintonD
and Chris J. HassellE
AJohnstone Centre, Charles Sturt University, Albury, NSW 2640, Australia.
Present address: 340 Ninks Road, St Andrews, Vic. 3761, Australia.
BRMB 4009, Cowes, Phillip Island, Vic. 3922, Australia.
CPhillip Island Nature Park, PO Box 97, Cowes, Phillip Island, Vic. 3922, Australia.
D165 Dalgety Rd, Beaumaris, Vic. 3193, Australia.
ETurnstone Nature Discovery, PO Box 3089, Broome, WA 6725, Australia.
FCorresponding author. Email: drogers@melbpc.org.au
Abstract.Two subspecies of the Gull-billed Tern (Sterna nilotica) occur along the coasts of north-western
Australia: the large, pale Australian subspecies macrotarsa, and a smaller, darker migratory subspecies from
northern Asia. On the basis of banding data we describe the measurements and moult strategies of both subspecies
in north-western Australia and identify the Asian migrants as subspecies affinis. Asian migrants have a predictable
plumage cycle including regular alternation between breeding and non-breeding plumage in adults. The moult
strategy of Australian macrotarsa is more varied and we argue it is adapted to exploit unpredictable breeding
opportunities. Plumage and structural characters described in this paper allow the two subspecies to be
distinguished in the field, and field observations demonstrate some broad ecological differences between them.
Adult affinis occur in Australia from August to April, with smaller numbers of immatures remaining during the dry
season; they are strictly coastal, occurring in highest abundance over intertidal flats near mangrove systems where
they pluck prey from the surface of mud while in flight. Subspecies macrotarsa uses the north-western Australian
coast as a non-breeding area, but it does so mainly during the dry season and also uses grasslands and inland
wetlands; unlike affinis, in Roebuck Bay it is regularly kleptoparasitic, stealing large crabs (Macropthalmus sp.)
from Whimbrels (Numenius phaeopus).
Gull-billed Terns in north-western Australia:
subspecies identification, moults and behavioural notes
www.publish.csiro.au/journals/emu
CSIRO PUBLISHING
D. I. Rogers et al.146 Emu
Davies 1996), it is possible that some Gull-billed Terns do so
too. (2) The type of subspecies affinis (Horsfield 1821) was
collected in Java. Although Peters (1934) and Gochfeld and
Burger (1999) regarded it as a doubtful form, most other
workers have considered it a valid subspecies that also occurs
in coastal south-eastern China (e.g. Dement’ev and
Gladkov 1951; Vaurie 1965; Meyer de Schauensee 1984;
C. S. Roselaar in Cramp 1985). More recent regional guides
have stated that the breeding range extends north to the
Bohai Sea (Étchécopar and Hüe 1978; Carey et al. 2001).
(3) The type specimen of subspecies addenda (Mathews
1912) was collected in China. Peters (1934) regarded it as a
valid form breeding on the Chinese coast and possibly
migrating to Indochina. Although most workers since have
considered addenda to be a synonym of affinis, Gochfeld
and Burger (1999) considered addenda a valid subspecies
with a breeding range including eastern China, Manchuria
and Transbaikalia and a wintering range mainly in South-east
Asia. The uncertainty about subspecies limits of Gull-billed
Terns in Asia is unlikely to be resolved without more data
from the breeding grounds. Nevertheless, an improved
knowledge of the morphology and plumages of the migrants
occurring in Australia may help to determine their precise
geographic origin.
Unlike the Asian subspecies, the validity of the distinctive
Australian subspecies macrotarsa has never been in doubt. It
has a strongly nomadic component to its movements, and can
breed opportunistically at any time of year if water levels in
inland wetlands are suitable (Bourke et al. 1973; Jaensch
1983; Davies 1988). Superimposed on this nomadism is a
seasonal component to its movements, with birds tending to
occur in southern or inland Australia during the austral
summer, and with many moving to the northern Australian
coast during the austral winter (Blakers et al. 1984; Davies
1988). This northward movement takes many Australian
Gull-billed Terns into the non-breeding range of migratory
Gull-billed Terns from Asia, not only along the northern
Australian coast but also in New Guinea (Mees 1982; Coates
1985), Mindanoa in the Philippines (Dickinson et al. 1991)
and possibly in some intervening islands of eastern Wallacea
(White and Bruce 1986; Coates et al. 1997).
Despite their broad overlap in range, few field observers
attempt to distinguish Australian and Asian Gull-billed Terns
and there have been no detailed comparative studies of the
two. Such studies might help to establish how subspeciation
developed in Gull-billed Terns; it is of particular interest that
birds of (presumably) the same genetic origin have diverged
to the point where some subspecies (e.g. nilotica, Cramp
1985) are long-distance migrants with tightly scheduled
annual cycles, while macrotarsa have sufficient plasticity in
their breeding and moult schedules to take advantage of the
unpredictable conditions of arid Australia. In this paper we
report on the moults, morphology and plumages of both sub-
species of Gull-billed Tern in Australia and attempt to iden-
tify the migratory subspecies from Asia. We also describe
ways in which Australian macrotarsa and Asian migrants can
be distinguished from one another in both the hand and the
field. Field observations made since we learned to make this
distinction allow us to document some broad differences in
the habitat choice, feeding behaviour and moult strategies of
the two subspecies.
Methods
Collection of banding data
Measurements and moult data were obtained from 629 birds captured
by the Australasian Wader Studies Group (AWSG) in north-western
Australia between 1982 and 2004. Most birds were cannon-netted at
high-tide roosts on sandy beaches, along Eighty Mile Beach (19°S,
121°E; n= 151) and the northern shores of Roebuck Bay (18°S, 122°E;
n= 474); four birds were mist-netted on adjacent Roebuck Plains. Few
or no breeding individuals are likely to have been caught, although
breeding colonies were established 3–30 km from the banding sites on
at least two occasions during the study period. Birds were banded,
weighed, and (if captured after 1992), given a yellow flag on the right
tarsus. The following measurements were taken regularly, using the
methods described in Marchant and Higgins (1990): wing (maximum
chord), exposed culmen and total head length (THL). Wing measure-
ments were discarded if the tenth primary (p10) was growing. Bill depth
(at the gonydeal angle) and tarsus were measured in a small number of
birds. Linear measurements of retraps were used in analyses only if the
variable in question had not been measured on first capture.
The degree of breeding plumage was recorded by estimating the
percentage of black in the cap; some observers made precise estimates,
and others placed individual birds into the most appropriate of the fol-
lowing categories: 0%, trace (treated as 10% in this paper), 25%, 50%,
75% or 100% breeding plumage. Primary moult was recorded using the
methods described in Marchant and Higgins (1990), with primaries
numbered from innermost (p1) to outermost (p10). In all birds, the con-
dition of each primary on one wing was scored as 1 (in pin), 2 (less than
one-third grown), 3 (between one- and two-thirds grown), 4 (more than
two-thirds grown, and still growing) or fully grown. Conventions for
scoring fully grown feathers changed over the years. Initially they were
simply scored as 0 (old) or 5 (new). This proved inadequate for the com-
plexities of moult in a species in which there are often three distinct
generations of fully grown feather in a single wing. After some trial and
error, a more detailed system of recording wear of individual primaries
was developed: they were scored as 0 (old), V (if very worn), 5 (new,
with no obvious abrasion of tips, or loss of silvery sheen from the dark
regions of the feather), 6 (if new and clearly a newer generation than
other primaries scored as 5 in the same wing), or 7 (if new and clearly
a newer generation than other primaries scored as 6 in the same wing).
In the last few years of the study, birds were identified at the banding
station as Australian or Asian, using a combination of plumage charac-
ters (described later in this paper) and general bulk (as a general rule of
thumb, Asian Gull-billed Terns can be held in one hand, while
Australian Gull-billed Terns require two). In addition we began record-
ing the intensity of crown moult by using a puff of breath to expose the
bases of the feathers; the abundance of growing crown feathers was
scored as 0 (absent), 1 (light), 2 (moderate) or 3 (heavy). When growing
crown feathers were found, their colour was described (black or white).
Birds were aged on plumage and moult characters; where possible,
we have named specific plumages and moults using the modifications
of the Humphrey and Parkes (1959) system of nomenclature recom-
mended by Howell et al. (2003). The consistent timing of moults of
Asian Gull-billed Terns made them easy to age; they were treated as
immature in this paper if they retained identifiable juvenile outer
Emu 147
primaries, and as adults if they did not. S .n . macrotarsa was consider-
ably more difficult to age. A few birds retained age-diagnostic juvenile
plumage of body, or still retained a complete set of juvenile primaries,
and these birds were treated as immatures. All other macrotarsa herein
are treated as adults, and it is likely that they include some birds in their
first plumage cycle that had started post-juvenile moult of primaries.
Field identification and behavioural observations
Most data were collected by D. I. Rogers. Initial familiarity with
plumages of known Australian macrotarsa and Asian migrants was
obtained through examination of banded birds in the hand, and of
known Australian macrotarsa and Indonesian affinis (from Java and
Bali) in museum collections; moult of the Indonesian specimens was
recorded using the protocols described above. Field descriptions and
behavioural notes were obtained opportunistically during shorebird
studies on the high-tide roosts and intertidal flats of Roebuck Bay
between 1997 and 2004. A large series of photographs, mainly obtained
by ‘digiscoping’ (Ingraham 2001) in June 2002 helped refine our under-
standing of some identification characters.
The distribution of shorebirds and other birds (including Gull-billed
Terns) on the intertidal flats of Roebuck Bay was mapped five times by
D. I. Rogers between 1997 and 2002. In brief, a grid of 200-m squares
was developed on the intertidal flats of northern Roebuck Bay, and
birds seen on the ground in each 200-m grid-square were counted. Birds
often moved around the intertidal flats in response to changes in tide
height. To minimise the effect that this had on mapped distributions,
mapping was done in a series of transects performed on receding tides;
the transects were timed so that the observer reached the sea-edge at low
water. A full description of the mapping methods is given by Rogers
(1999).
Information on the migratory departures of Asian Gull-billed Terns
from Roebuck Bay was obtained in the course of regular migration
watches by the AWSG and Broome Bird Observatory. Daily migration
watches from 1600 to 1800 hours were made through March and April
over a period of 12 years. Observations were made from south-facing
beaches, especially one next to Broome Bird Observatory; shorebirds
setting off on northwards migration thus passed over the observers and
could be identified and counted. The time at which each flock departed,
shape of flock (e.g. echelon or V) and direction in which it was heading
were also recorded. A number of distinctive behaviours and vocalisa-
tions (cf. Lane and Jessop 1985; Piersma et al. 1990) helped to distin-
guish migratory departures from other flights by shorebirds.
Analysis
The banding data used in this study were obtained over a 20-year period
by many different observers. Although these observers were consistent
in their measuring methods, description of the amount of breeding
plumage and recording of the location of growing primaries, it was only
in the later years of the study that birds were routinely identified to sub-
species in the hand, and that consistent ways of describing primary wear
were found. As we did not wish to discard all previous data, we made
post-hoc subspecific identifications using the following approach.
Examination of the histograms of bill length and THL suggested that
each had a bimodal distribution. The SHEBA programs of Rogers
(1995) are designed to separate the component normal distributions of
such double-humped distributions. We used SHEBA to estimate the
measurement parameters (mean, standard deviation and n) of bill length
and THL for birds from the larger and smaller cohorts. The differences
in size between Asian and Australian birds far exceeded the size dif-
ference owing to sexual dimorphism, which was therefore ignored at
this stage of the process. The estimates of size of the larger and smaller
cohorts were used to calculate identification criteria: e.g. cut-off values
below which the probability of a bird being from the smaller subspecies
was at least 95%, and above which the probability of a bird being from
the larger subspecies was at least 95%. As a starting point, we treated
birds as correctly identified if: (1) they had been identified to sub-
species from photographs or at the banding site; or (2) if they could be
identified with at least 99% confidence on bill length and at least 90%
confidence on THL; or (3) if they could be identified with at least 99%
confidence on THL and at least 90% confidence on bill length.
We identified 515 birds on these criteria, including 105 of the
smaller subspecies. We then examined their moult records. There was a
characteristic ‘signature’ to the moult strategy of Asian birds (see
Results), which showed very consistent timing of primary moult, and
timing of moult to breeding plumage. Remaining birds were assigned to
macrotarsa if they did not show this moult signature. Individuals that
did show the ‘Asian’ moult signature could in theory have been
macrotarsa that happened to be moulting on a similar schedule to Asian
birds, so we only assigned such birds as Asian if their bill length or THL
measurements supported this identification with a probability of at least
90%. There remained 29 birds that we could not identify with certainty
using moult characters. These were assigned to subspecies simply on
the basis of measurements: at least one of wing or weight fitted only
into the range of one of the two subspecies (judging by the parameters
developed from the birds identified by the plumage characters, moult
criteria or 99% measurement criteria described above). If wing or
weight were not diagnostic, birds were assigned to the most probable
subspecies on the basis of bill length and THL. Differences in size
between the birds identified to subspecies on plumage characters and
those identified by the other criteria were negligible, indicating that the
identifications made on the basis of measurements and moult signature
were adequate.
Results
Measurements, subspecies identification and
band recoveries
The measurements obtained during the banding study from
adults and immatures are summarised in Tables 1 and 2,
respectively. Within both age categories, Asian migrants
were significantly smaller than Australian macrotarsa for all
measurements (two-tailed t-tests; P< 0.02) except length of
wing (in fresh plumage) and bill depths of immatures, for
which sample sizes were small. In subspecies macrotarsa,
adults had significantly greater bill lengths, THL and
weights than immatures (two-tailed t-tests; P< 0.01). In
Asian migrants, adults had significantly longer wings (when
worn) than immatures (two-tailed t-tests; P< 0.01). No other
age-related differences in size were found, though they might
have been obscured by small sample sizes for immatures.
Wing length was affected by wear: in adults of both sub-
species, the wing of birds in which p10 was fresh was signif-
icantly longer than wing of birds with worn p10 (two-tailed
t-tests, P< 0.01).
Examination of histograms of bill length and THL sug-
gested that within both subspecies, these measurements had
a bimodal distribution, presumably caused by sexual size
dimorphism. Males are the larger sex in Gull-billed Terns
(C. S. Roselaar in Cramp 1985; Higgins and Davies 1996).
Estimates of the THL of males and females for each sub-
species in north-western Australia were made with the
SHEBA programs, assuming the co-efficient of variation to
be equal in each sex. Using the sexes found on this basis, the
Gull-billed Terns in North-western Australia
D. I. Rogers et al.148 Emu
bill and wing lengths of both subspecies could also be calcu-
lated (Table 3). The measurements of Asian migrants corre-
spond well with previously published measurements of sexed
museum skins of subspecies affinis from China and
Indonesia (Table 3), and are significantly smaller than those
available for nominate nilotica (two-tailed t-tests, P< 0.01).
Weights of Gull-billed Terns in north-western Australia
are presented by month in Fig. 1. Although samples are too
small for significance testing, there appears to be a tendency
for adults of the Asian subspecies to be heaviest in April, just
before northwards migration. Average weight in April was
205.6 ± 21.3 g (range 168–240 g), cf. 156.7 ± 7.4 g (range
149–168) in January, suggesting pre-migratory mass-gain
may have been over 50 g. April was the only month in which
adults were significantly heavier than immatures (which do
not migrate; see later). No age-related differences in weight
were found in subspecies macrotarsa, but the data plotted in
Fig. 1 suggest this subspecies is heaviest in August and
September, late in the dry season.
There were no band recoveries or flag resightings of
Asian migrants outside the study area. Three Australian
macrotarsa banded on Eighty Mile Beach in August and
September 1998 were found, dying or dead, at breeding
colonies at Mandora Marsh, 30 km from the banding site, in
June 1999 (n= 2) and October 1999. Many Gull-billed Terns
and Australian Pelican (Pelecanus conspicillatus) were
dying at the site then, apparently from botulism. A yellow-
flagged macrotarsa was also resighted at Lake Gregory
(20°10′S, 127°30′E), 600–700 km from the banding site on
30 December 1996.
Moults of Asian migrants
Asian migrants had a moult strategy typical of migratory
Sterna terns, including nominate nilotica (cf. C. S. Roselaar
in Cramp 1985), making the identification of their moults
straightforward (Fig. 2). Fig. 2Acompares the average per-
centage of definitive alternate (breeding) plumage in the
crown found by month in north-western Australia and in
museum specimens from Indonesia. The two populations
correspond closely. Accordingly, breeding plumage data
from both are combined in Fig. 2B, which shows they spend
most of the austral summer in a white-capped, definitive
basic (non-breeding) plumage. The cap begins to change
colour in February and was fully black in most birds sampled
in April. We have no data from the breeding grounds; breed-
ing plumage is probably held through most of the breeding
season in the northern hemisphere, as it is in nominate
nilotica (C. S. Roselaar in Cramp 1985; Olsen and Larsson
1995). Little retained breeding plumage was found in
recently returned adults from August to October, suggesting
that most pre-basic moult of head feathers occurs before or
during southwards migration.
The timing of the moult of primaries of Asian Gull-billed
Ter ns in north-western Australia and subspecies affinis in
Tab le 2. Measurements of live immature Gull-billed Terns from north-western Australia
Linear measurements in mm; weight in g. Wing (f) = Wing length of birds with fresh p10; Wing (w) = wing length of birds with
worn p10. Asian migrants include all birds with p10 retained from juvenile plumage. S. n. macrotarsa only treated as immature if they
retained some juvenile body plumage. Differences between subspecies were significant (P< 0.02) for all measurements except wing (f)
and bill depth
Parameter S. n. macrotarsa Asian migrants
Mean s.d. nRange Mean s.d. nRange
Wing (f) 327.7 10.5 3 317–338 301.3 9.02 3 317–328
Wing (w) 321.8 13.0 5 306–337 275.6 15.9 8 248–296
Bill length 43.4 1.84 9 40.8–45.9 38.0 2.13 25 34.5–41.5
THL 91.1 2.54 9 87.8–94.9 82.1 2.50 25 76.2–88.2
Bill depth 12.8 3.32 2 10.4–15.1 10.05 0.21 4 9.8–10.3
Tarsus length –––– 33.8 2.72 4 31.1–37.6
Weight 229.4 31.3 8 187–281 169.5 12.6 25 146–199
Tab le 1. Measurements of live adult Gull-billed Terns from north-western Australia
Linear measurements in mm, weight in g. Wing (f) = Wing length of birds with fresh p10; Wing (w) = wing length of birds with
worn p10. Differences between subspecies were significant (P< 0.02) for all measurements
Parameter S. n. macrotarsa Asian migrants
Mean s.d. nRange Mean s.d. nRange
Wing (f) 338.6 9.91 149 305–367 312.7 9.21 24 292–327
Wing (w) 333.2 14.4 240 283–367 298.5 9.79 57 271–320
Bill length 45.6 2.48 420 38.3–51.6 38.1 2.35 103 32.8–43.4
THL 93.5 3.16 414 80.4–102 83.3 2.90 103 76.8–90.0
Bill depth 11.7 0.69 19 10.4–13.3 10.4 0.54 16 9.5–11.4
Tarsus length 37.4 3.21 11 32.3–43.7 33.9 3.03 12 30.8–40.1
Weight 253.3 22.0 489 204–369 182.1 19.9 103 141–240
Emu 149
Indonesia also corresponded closely (Fig. 2C). In both
regions, a tightly scheduled, complete pre-basic primary
moult is centred in the austral summer. Of 34 adult Asian
birds captured on the non-breeding grounds from August to
October, 14 had suspended primary moult after replacing
3–5 inner primaries. Nominate nilotica begins primary
moult when on or near the breeding grounds and then sus-
pends moult during southwards migration (Cramp 1985), a
scenario consistent with most of our data for Asian birds. We
also caught four individuals in the August–October period
that had yet to begin primary moult.
Once the pre-basic moult of primaries of adults began or
resumed in the non-breeding areas, it continued uninter-
rupted during the austral summer (Fig. 2D), with the last
Gull-billed Terns in North-western Australia
Tab le 3. Measurements of male and female adult Gull-billed Terns
Measurements (in mm) are presented in the format mean (s.d.; n). Sources of information:
(1) This studyA; (2) C. S. Roselaar in Cramp (1985); (3) Piechocki (1968). Except where stated,
all samples from north-western Australia
Subspecies (and location) Males Females Source
Wing length
S. n. macrotarsa 335.4 (13.02; 232) 330.0 (13.46; 117) 1
S. n. macrotarsa – fresh 341.1 (9.08; 79) 335.3 (8.69; 40) 1
S. n. macrotarsa – worn 333.7 (13.40; 133) 328.5 (14.44; 67) 1
Asian migrants 305.8 (10.45; 45) 298.8 (4.42; 26) 1
Asian migrants – fresh 313.7 (8.95; 16) 313.7 (8.16; 6) 1
Asian migrants – worn 301.7 (8.64; 29) 293.9 (3.63; 20) 1
S. n. affinis (Indonesia) 312 (7.12; 10) 295 (12.5; 8) 2
S. n. nilotica (Europe to Africa) 326 (7.36; 35) 319 (7.41; 24) 2
S. n. ssp (Mongolia) 313.0 (10.84; 5) 301.0 (15.56; 2) 3
Bill length
S. n. macrotarsa 46.6 (2.17; 267) 44.0 (2.05; 132) 1
Asian migrants 38.6 (1.73; 71) 35.8 (1.61; 37) 1
S. n. affinis (Indonesia) 38.4 (1.12; 12) 35.5 (1.42; 10) 2
S. n. nilotica (Europe to Africa) 39.8 (1.44; 36) 37.7 (1.37; 28) 2
THL
S. n. macrotarsa 94.9 (2.02; 267) 90.5 (1.93; 132) 1
Asian migrants 84.4 (1.58; 71) 79.9 (1.50; 37) 1
ASexing of north-western Australian birds based on statistical separation of the sexes on THL using the
SHEBA programs of Rogers (1995).
Fig. 1. Mean weights of Gull-billed Terns by
month. Sample sizes are given next to means;
error bars depict one standard deviation.
D. I. Rogers et al.150 Emu
(outermost) primary being moulted in March or the first half
of April. Duration of this primary moult in individual birds
must therefore be at least 150 days, and is probably 1–2
months longer. Before this wave of primary moult was com-
pleted, most individuals began a pre-alternate wave of
primary moult in January or February. This moult started at
p1, and 3–6 inner primaries were replaced before the pre-
alternate moult wave was interrupted in April.
In the absence of moult data from or near the breeding
grounds, it is difficult to prove whether the pre-alternate
moult wave observed in February–April was suspended
(resuming at the point of interruption after breeding) or
genuinely arrested. The number of primaries replaced in
February to March (4.2 ± 0.7, n= 12) was similar to the
number of moulted primaries observed in birds with sus-
pended moult captured on the non-breeding grounds from
August to October (4.0 ± 0.7, n= 9). In the August–October
samples, the outer primaries were noticeably worn (hence the
shorter wing length in birds with worn primaries, Table 1),
suggesting they were unlikely to have been replaced after
breeding. We therefore consider the pre-alternate moult of
adult Asian Gull-billed Terns to be arrested, i.e. the outer
primaries are replaced once per plumage cycle, and the inner
primaries twice.
In the first plumage cycle, pre-formative (post-juvenile)
moult of body feathers began after the onset of southwards
Fig. 2. Moults of adult Asian migrants. (A) Average percentage of breeding plumage ± s.e. (B) Number of birds caught with black, white
or mottled crown. (C) Average outermost growing primary ± s.e. (outermost primary = p10). (D) Number of birds caught with one, two
or three active moult loci in primaries.
Aug Sep Oct Nov Dec Jan Feb Mar Apr May
Aug Sep Oct Nov Dec Jan Feb Mar Apr May
A
B
C
D
13
11 11
1
11
1
3333
3
3
27
21
4
4
4
21 2 2
2
2
2
2
8
8
19
14
14
6
15
1
100
80
60
40
20
0
25
20
15
10
5
0
p10
p8
p6
p4
p2
0
0
30
25
20
15
10
5
Location
Asia
Asia Prealt.
Asia Prebasic
NWA
Crown colour
Black
Mottled
White
NWA Prealt.
NWA Prebasic
No. of loci
Two
One
None
No. of birds Outermost
growing primary No. of birds % Breeding plumage
Moult wave
Emu 151
migration, judging by a single specimen in complete juvenile
plumage collected in Java on 9 September; another collected
in Java on 22 October had moulted 95% of its juvenile body
plumage, though it still retained juvenile tertials and
remiges. By November, all first-year Asian birds captured
had attained formative (first non-breeding) head and body
plumage, and thereafter both Australian and Indonesian birds
were white-capped throughout the first moult cycle. We do
not know whether a first pre-alternate body moult occurs,
but if it does, it produces no obvious change in appearance.
Unfortunately few data on intensity of crown moult were col-
lected from immatures during the austral autumn and winter.
Two birds in June and one in August showed no crown moult,
suggesting there may be no pre-alternate body moult; on the
other hand, the crown plumage of year-old birds in August
did not look particularly worn.
As in adults, the timing of primary moult of Asian
migrants in their first cycle was similar in north-western
Australia and Indonesia, and it showed little variation
(Fig. 3A). A pre-formative moult wave starting at p1 and
moving outwards began on the non-breeding grounds in
November or December, and proceeded steadily until about
September or October when the birds were 14–15 months
old. Although this primary moult lasted 8–9 months, it was
seldom suspended (Fig. 3B). Second pre-basic primary
moult sometimes started before the pre-formative moult
wave had finished, so from June to August some of the
immatures sampled had two moult loci in the primaries.
Second pre-basic moult of the primaries began at about the
same time as the complete pre-basic primary moults of
adults, and when it was well advanced, second year and older
birds became indistinguishable.
Moults of Australian macrotarsa
The moult strategy of the Australian subspecies macrotarsa
differed greatly from that of Asian migrants, and it was dif-
ficult to identify individual moults. Johnstone and Storr
(1998) stated that adult macrotarsa do not develop a non-
breeding plumage. This is not consistent with our data on
crown plumage: (1) of 13 birds known from retrap history to
be at least 2–6 years old, only six had complete breeding
plumage; five had 50–95% breeding plumage and one had
no breeding plumage at all; (2) 32 individuals had more
breeding plumage on recapture than when originally caught,
10 did not change appearance, and 20 had less breeding
plumage on recapture; and (3) many individuals were noted
to be replacing black feathers in the crown with white ones
(further details below). Clearly, some seasonal change in
plumage appearance can occur in macrotarsa. In combi-
nation with their variable annual cycle, this makes them dif-
ficult to age and we were only able to age a few first-year
individuals that retained some juvenile body plumage. Moult
data from all other macrotarsa are included in Fig. 4 and
while adults probably predominate in these graphs, it is
likely some younger birds are also included.
The proportion of birds with a wholly black crown was
lowest in the austral winter and increased during the austral
spring (Fig. 4A); few birds were caught in the middle of the
austral summer, a period in which our (unquantified) impres-
sion from bird-watching is that most macrotarsa have com-
Gull-billed Terns in North-western Australia
Fig. 3. Moults of immature Asian migrants. (A) Average outermost growing primary ± s.e. (outermost primary = p10). (B) Number
of birds caught with one, two or three active moult loci in primaries.
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
A
B
11
1
3
4
2
Asia Prealt.
Asia Prebasic
NWA Prealt.
NWA Prebasic
No. of loci
Two
One
None
No. of birds
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
Moult wave
Outermost
growing primary
p10
p8
p6
p4
p2
p0
10
8
6
4
2
0
244
1
1
1
1
1
1
3
4
D. I. Rogers et al.152 Emu
plete breeding plumage. In all adequately sampled months,
many macrotarsa had complete breeding plumage and many
others had a mottled crown (Fig. 4A). In all months the
appearance of birds with mottled crowns varied consider-
ably, with a continuum ranging from birds with white crowns
and only a trace of black mottling, to black-crowned birds
with only traces of white. Although some ‘mottled’ birds
might have been in a transitional state (i.e. moulting from
breeding to non-breeding plumage or vice versa), many had
no active moult of crown feathers (Table 4). Moreover, there
was little temporal consistency in the colour of growing
crown feathers; emergent black feathers and emergent white
feathers were recorded between June and November, with
both conditions often being seen in the same catch. Emergent
black feathers were more likely to be seen in birds with
extensive breeding plumage (Table 4).
Figure 4B, Csummarise the primary moult of macrotarsa
in north-western Australia. We believe each moult locus
found in macrotarsa represents the front of a primary moult
wave that was moving outwards, because (1) if a moult locus
involved more than one growing primary, the outermost was
always the youngest feather; and (2) in general, the feathers
on the proximal side of the locus were fresher than those on
the distal side.
It was difficult to detect any relationship between primary
moult and time. In most sampled months, the outermost
growing primary ranged from p1 to p10, and there was often
an additional locus of growing primaries inside that
(Fig. 4B), in which moult began at p1 at about the time that
the outer moult wave reached p5–p6. The inner wave reached
~p4–p7 by the time the outer wave reached p10.
Distinguishing pre-basic from pre-alternate waves of
primary moult was problematic. We made an attempt by
assuming that if there were two loci, the innermost was pre-
alternate, and if there was one locus, it was pre-basic, unless
the primaries outside that were specifically noted to be fresh
(in which case we assumed that the pre-basic moult wave had
been completed and remaining active moult was pre-
alternate). A plot of the average outermost growing primary
of each feather generation, made by following the above
Fig. 4. Moults of apparently adult S. n. macrotarsa. (A) Number of birds caught with black, white or mottled crown. (B) Lines
represent average outermost growing primary ± s.e. (outermost primary = p10); numbers represent number growing each primary as
part of outer or inner wave of moult. (C) Number of birds caught with one, two or three active moult loci in primaries.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
A
B
C
2 40 23 189 52 48 2 107 3
100
80
60
40
20
0
p10
p8
p6
p4
p2
Prebasic
Crown colour
Black
Mottled
White
Prealt.
Presupp.
No. of loci
Two
One
None
% of active Primary number No. of birds
Moult wave
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
1 6 41 24 193 66 48 2 107 3
p9
p7
p5
p3
p1
100
80
60
40
20
0
moult loci
Three
Emu 153
assumptions, is shown in Fig. 4B. The lack of variation in
calculated stage of pre-basic primary moult suggests our
basis for identifying moult was imperfect, especially in
November catches. Nevertheless, the plot suggests that a pre-
alternate wave proceeded slowly through the austral autumn
and early winter (slightly earlier than a similar increase in the
amount of breeding plumage, Fig. 4A). Of birds that had full
black caps and lacked active primary moult, 20 were
recorded as having no moult contrast (this may have been an
overestimate caused by failure of some processors to detect
subtle differences in wear), while in 15 there was a moult
contrast between fresh inner and more worn outer primaries.
The average number of fresher primaries in such birds was
5.1 (s.d. = 1.98, range 2–9, n= 15). Of 66 birds with only one
moult locus between p1 and p5, 30 had a wear contrast
within the unmoulted outer primaries, with the outer feathers
being older. In contrast, of 161 birds with one moult locus
between p6 and p10, only 31 showed any contrasts between
worn and fresh primaries between p1 and p5. These data
suggest the pre-alternate moult wave was often arrested,
i.e. not resuming at the point of interruption. In this respect
it thus resembled the pre-alternate moult of affinis.
The number of active moult loci in the primaries in
macrotarsa was much less consistent than in Asian migrants,
which typically had one locus except in short periods when
different moult cycles overlapped. In contrast, many
macrotarsa had interrupted primary moult (hence having no
active loci) in nearly all sampled months, and many had two
loci (Fig. 4D) in all sampled months. A few individuals had
a third moult locus in the primaries. There was substantial
variation in the number of moult loci, the position of the
outermost growing primary, and the amount of breeding
plumage, within single catches, e.g. the November data in
Fig. 4 were all collected in 2003. Our data were not well
suited to testing whether or not there were significant dif-
ferences between timing of moult in different years.
However, this is likely to occur, given that in three years
when unusually wet conditions allowed birds to nest at the
nearby Mandora Marshes (30 km from Eighty Mile Beach,
May–October 1999 and May 2004) or Anna Plains (3 km
from Eighty Mile Beach, May 2000), adults observed and
photographed at these colonies had full breeding plumage
and no active primary moult. They had probably interrupted
moult to breed.
Field identification
Given adequate viewing conditions, S. n. macrotarsa and
Asian migrants in northern Australia can be distinguished
reliably in the field. Field identification characters are
described by Rogers (2004) and are only summarised briefly
here. Key distinguishing features are:
(1) The larger size of S. n. macrotarsa is obvious when Asian
migrants are available for direct comparison.
(2) The bill of S. n. macrotarsa has a decurved culmen and
weak gonydeal angle relatively close to the bill tip, and
thus appears more drooped than the very straight bill of
Asian migrants, in which the culmen is almost straight
and the prominent gonydeal angle is closer to the base of
the bill.
(3) The upperparts of S. n. macrotarsa are paler grey than in
Asian migrants, creating a field impression of silvery
upperparts that grade to wholly white rump and tail. In
Asian migrants in flight, there is generally an obvious
contrast between white uppertail coverts and light to
pale-grey back and tail; the resultant white patch centred
on the uppertail coverts can be discerned at considerable
distances. When Asian migrants spread their tail in flight,
the white outer tail-feathers appear distinctly paler than
the pale-grey central feathers, an effect not seen in the
wholly white-tailed macrotarsa.
(4) In non-breeding plumage, S. n. macrotarsa develops a
large black oval patch encircling the ear-coverts and eye.
Asian migrants have less black on the face; typically
there is a small black patch around the eye and a small
dark spot on the ear coverts; in some individuals these
spots meet, but there is a strong constriction between
them. In addition Asian migrants have a conspicuous
white lower eyelid that accentuates the division between
dark eye- and ear-patches. Although not completely diag-
nostic, S. n. macrotarsa usually has some black mottling
in the crown in non-breeding plumage, while the cap of
Asian migrants in non-breeding plumage is usually
unsullied white.
Gull-billed Terns in North-western Australia
Tab le 4. Appearance of crown and moult condition of S. n. macrotarsa
The number of birds in each category of cap colour is shown. The final two columns also show the months in which in which
incoming black or white feathers were recorded
% breeding Total No crown Active crown Incoming black feathers Incoming white feathers
plumage examined moult moult
041132618: Jul., Sep., Nov. 10: Jul., Aug., Nov.
Trace 14 7 5 3: Jul., Nov. 3: Sep., Nov.
25 29 13 15 7: Jun.–Sep., Nov. 6: Jun.–Jul., Sep.
50 69 43 22 20: Jun.–Jul., Sep., Nov. 5: Jun.–Jul.
75 54 25 22 21: Jun.–Sep., Nov. 0
100 110 46 64 58: Jun.–Sep., Nov., Apr. 0
D. I. Rogers et al.154 Emu
(5) With their variable timing of breeding and moult,
S. n. macrotarsa can be seen in any plumage type at any
time of year. In contrast, some plumages of Asian
migrants are only seen at some times of year: adult
breeding plumage (with complete black cap) is only
likely to be seen in Australia from March to the start of
May; if complete juvenile plumage (with buff tips and
dark subterminal markings to body feathers) ever occurs
in Australia, it will only be in September and early
October; formative plumage (with non-breeding head
and body plumage, but with retained juvenile remiges
forming uniformly dark primary tips at rest, and a con-
tinuous dark bar across the secondaries in flight) only
occurs from October to January.
Behavioural notes
Nearly all Gull-billed Terns were captured at high-tide roosts
on sandy beaches, usually in mixed flocks with other species
of tern and shorebirds. A break-up of catch totals by date
(Table 5) indicates that Asian migrants were most likely to be
captured from October to April but smaller numbers of birds
stayed in Australia from May to September. The reverse
occurred in macrotarsa, which was caught most often in the
dry season from June to November. There was also a dif-
ference between sites: only 11.7% of the 471 Gull-billed
Ter ns caught in Roebuck Bay were Asian migrants, but at
Eighty Mile Beach Asian migrants comprised 47.4% of the
152 Gull-billed Terns caught.
The frequent occurrence of Gull-billed Terns on high-tide
roosts at Eighty Mile Beach and Roebuck Bay suggests that
they have intertidal feeding habitats, an interpretation con-
sistent with our field observations. Single Gull-billed Terns
were frequently seen standing on the mudflats. Such stand-
ing birds were almost invariably macrotarsa, and they were
generally close to Whimbrels (Numenius phaeopus) (Rogers
1999). The correlation between Whimbrel and S. n. macro-
tarsa distribution was statistically significant in all but one
survey (Table 6), and was indeed even stronger than these
χ2tests suggest, as in the few cases where S. n. macrotarsa
were found in grid-squares without Whimbrels, there were
Whimbrels in the adjacent grid-squares. Field observations
indicated that this was a kleptoparasitic association. The
Whimbrels were hunting large sentinel crabs
(Macropthalmus sp.) and mantis shrimps, usually seeing
them on the surface, making chase and grabbing them from
near or in the mouth of their burrows. Long handling times
(often a minute or more) were associated with these large
prey items. Each macrotarsa kept at least one Whimbrel
under careful observation, waiting until the Whimbrel caught
a crab or shrimp, and then immediately flying in to steal it.
Fights over possession of a crab were frequent, the Gull-
billed Terns initially flying in quietly to attempt a clean
snatch, and if this did not work, persistently charging the
Whimbrel with loud calls; physical contact was rare, but
direct buffeting was sometimes seen. Some attacks made by
Gull-billed Terns were successful; sometimes they would
manage to steal an entire crab, sometimes the Whimbrel got
away with it, and more often, the crab would be broken, with
one bird eating the carapace while the other made do with
legs that had broken off.
Kleptoparasitism of Whimbrels was not the only feeding
method used by macrotarsa around Roebuck Bay. They were
Tab le 5. Numbers of each subspecies caught by month on Eighty Mile Beach and Roebuck Bay
Jun. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Total
Asian migrants
Roebuck Bay 2 17 1 3 0 4 2 5 3 0 14 4 55
Eighty Mile Beach 6 2 0 15 1 0 8 1 1 28 10 0 72
Total 8 19 1 18 1 4 10 6 4 28 24 4 127
S. n. macrotarsa
Roebuck Bay 0 0 5 35 0 21 176 44 29 0 103 3 416
Eighty Mile Beach 1 0 0 7 0 3 18 27 19 3 2 0 80
Total 1 0 5 42 0 24 194 71 48 3 105 3 496
Tab le 6. Co-occurrence of Whimbrels with subspecies macrotarsa on the intertidal flats of northern Roebuck Bay
The number of 200-m grid-squares containing at least one Whimbrel and one macrotarsa is shown. The null hypothesis
for the χ2tests was that the distributions of Whimbrels and Gull-billed Terns were independent of one another
Survey date Both species Whimbrels S. n. macrotarsa Both species χ2P(d.f. = 1)
absent only only present
Oct. 1997 74 92 0 19 12.32 <0.01
Feb. 1998 63 45 6 23 11.50 <0.01
Aug. 1998 151 29 7 6 5.49 <0.05
Feb. 2000 101 59 2 5 2.08 >0.05
Jun 2002 130 5 15 29 79.46 <0.01
Emu 155
also often encountered on the grasslands and freshwater wet-
lands of Roebuck Plains, flying slowly at 5–10 m above the
ground and occasionally swooping low to pluck a prey item
from the surface without landing; this would then be swal-
lowed in flight. Usually we were unable to identify the prey
taken, but it included frogs and small lizards. Small lizards,
probably caught by this method, were often fed to chicks at
Anna Plains in 2000 (C. D. T. Minton, unpublished observa-
tions), and at the Mandora Marshes in 2004 (J. van de Kam,
personal communication). The same feeding method was
occasionally seen on the intertidal flats, but was not observed
as often as kleptoparasitism.
Kleptoparasitism did not appear to be an important
feeding technique for Asian migrants. We never saw Asian
migrants attempt to rob Whimbrels; the only record of
kleptoparasitism that we are aware of is of a single bird seen
attacking a Grey-tailed Tattler (C. J. Hassell, unpublished
observation). Of 865 Gull-billed Terns mapped standing on
the intertidal flats of northern Roebuck Bay, only four were
Asian migrants. However, Asian migrants were often seen
flying over the intertidal flats, feeding by making low swoops
to pluck prey from the surface without landing. They were
widely distributed over the intertidal flats and were usually
seen foraging alone, but loose concentrations of up to 30
birds were seen feeding near mangroves on the southern
shores of Roebuck Bay, and in large unvegetated areas of
mud within the mangrove systems in the north-east of the
bay. In the latter region they were seen catching female
Flame Fiddler Crabs (Uca flammula). Small unidentified
crabs were also taken from the surface in other parts of the
bay. We never found Asian migrants on grasslands or inland
wetlands near Eighty Mile Beach or Roebuck Bay. It is pos-
sible we overlooked a small number, as it was sometimes dif-
ficult to get close to Gull-billed Terns in these habitats, but
macrotarsa certainly predominated.
Asian migrants were often seen setting off on northwards
migration from Roebuck Bay in March and April. Observed
departure behaviour was closely similar to that of the shore-
birds of Roebuck Bay (described in Lane and Jessop 1985;
also see Piersma et al. 1990). In late afternoon and early
evening, flocks became increasingly excited, flying high
with distinctive yelping ‘kee-yep’calls, which we do not
think are made by Gull-billed Terns at other times; this call
has been recorded and published by Stewart (1999). Flocks
often alighted several times before eventually settling into
level flight and disappearing to the north in line-formation
(48% of the 25 departures for which flock formation was
recorded), in V-formation (16%) or in a loose association
(36%). Average size of flocks was 12.4 birds (s.d. = 9.75,
range 1–50, n= 87). Of 1051 Gull-billed Terns observed
departing, 14.4% left in the last week of March, 36.6% in the
first week of April and 29.9% in the second week of April;
the earliest observed departure was 22 March, and the latest
8 May. Usually, departing Gull-billed Terns were observed in
single-species flocks, but they were occasionally seen
leaving in mixed flocks with Bar-tailed Godwits (Limosa
lapponica), Whimbrels or Common Greenshank (Tringa
nebularia).
Discussion
Asian Gull-billed Terns were common in our north-western
Australian samples, comprising 20.4% of the 623 Gull-billed
Ter ns captured. The proportion of Asian migrants sampled
may have been biased by capture date and differences
between the subspecies in roosting behaviour, so this cannot
be regarded as a precise estimate of the proportion of Asian
to Australian Gull-billed Terns in the region. Nevertheless, it
is clear that Asian migrants are more than a vagrant to north-
western Australia. Elsewhere in Australia, we are aware of
records of Asian Gull-billed Terns from Derby in the
Kimberley (D. J. James, personal communication), from the
Darwin coast and the estuaries of the Alligator Rivers in the
Northern Territory (McKean 1981; Schodde 1991) and
Torres Strait (M. J. Carter, personal communication). There
are also records of vagrants from Whyte Island in south-
eastern Queensland (two individuals seen several times
during December 1990 and January 1991 (C. Corben,
personal communication), a single bird at Stockton Bridge,
Newcastle, NSW, on 22 February 1991 (D. Hobcroft,
personal communication) and three birds from Boundary
Road Wetland, Carrum Downs, Vic., on 13 March 2005
(M. J. Carter and P. S. Lansley, personal communication)).
Given that few bird-watchers have attempted to distinguish
Asian migrants from Australian macrotarsa, it is possible
that the northern Australian coast is a major non-breeding
area for Asian migrants.
The subspecific identity of the Asian migrants in
Australia has not been proved beyond all doubt, but the most
likely identification is S. n. affinis. The small, heavy-billed,
dark-backed Asian migrants caught in north-western
Australia are significantly smaller than typical nominate
nilotica, and indistinguishable from specimens collected in
Java (C. S. Roselaar in Cramp 1985; D. I. Rogers, this study),
the type locality of affinis. In both Indonesia and north-
western Australia, the tightly scheduled timing of moults is
indicative of regular migration to breeding grounds where
nesting occurs during the boreal summer. We do not know
where these breeding grounds are, although the observation
that north-western Australian birds sometimes set off on
northwards migration with Bar-tailed Godwits, which are
bound for the Yellow Sea (Barter 2002), suggests a far-
eastern destination. The breeding grounds may lie in
southern or eastern China, where the type specimen of
addenda Mathews 1912 was collected. Peters (1934) evi-
dently considered the Gull-billed Terns of southern and
eastern China to be locally resident, and this was presumably
the basis of his decision to recognise addenda. However, it
has since been shown that many or all Gull-billed Terns in
Gull-billed Terns in North-western Australia
D. I. Rogers et al.156 Emu
this region are summer visitors or passage migrants (Carey
et al. 2001), with a non-breeding range to the south.
C. S. Roselaar in Cramp 1985, personal communication)
could find no morphological differences between Indonesian
and Chinese birds, nor are we aware of any published evi-
dence to suggest that non-breeding birds from the inter-
vening regions (e.g. Thailand, the Malay Peninsula) differ
from Indonesian migrants. It therefore seems most likely that
the non-breeding range of birds breeding in southern or
eastern China includes the type locality of affinis. If this is so
then addenda should be relegated to synonomy, as affinis is
the older name.
Alternatively, the non-breeding migrants of Indonesia and
northern Australia may come from farther north, in
Manchuria or Mongolia. These regions are supposedly
inhabited by nominate nilotica, but this interpretation seems
to have been made rather tentatively (e.g. see Vaurie 1965;
Étchécopar and Hüe 1978). In the only measurements we
have found from Mongolia (Piechocki 1968), wing lengths
corresponded well with those of Asian migrants in Australia
and were shorter than those of nominate European S. n. nilo-
tica. We argue that if the Gull-billed Terns of Manchuria and
Mongolia are similar in size to nominate nilotica found
farther west, then they are too big to include migrants from
Australia and Indonesia; if they are the same size as
Australian and Indonesian birds, then it is difficult to see
why they should be classified as nominate nilotica, a sub-
species that differs from affinis mainly in size (C. S. Roselaar
in Cramp 1985). There remains the caveat that we cannot rule
out intergradation between nominate nilotica and affinis.
However, in the absence of adequate documentation of such
hypothetical intermediates, it is prudent to continue to accept
affinis as a valid subspecies.
Migratory departures of affinis were often observed from
Roebuck Bay. The occasional association of these departing
birds with migratory shorebirds, their pre-migratory mass-
gain, and the apparent lack of records at sea (Higgins and
Davies 1996), all suggest that the affinis of north-western
Australia migrate in prolonged non-stop flights rather than
feeding at sea on their way north. We are unable to estimate
their potential flight-range, as our small April samples may
not represent the final departure mass. However, as the
average weight in April was 10.9–31.4% greater than
average weights in November to January respectively, it is
likely that they are capable of a direct flight of several
hundred or thousand kilometres. For comparison, Great Knot
(Calidris tenuirostris) that fly directly to the Yellow Sea from
north-western Australia (at least 5400 km) undergo pre-
migratory mass-gain of ~67% (Higgins and Davies 1996;
Battley et al. 2000).
In Australia all our records of affinis are coastal and all
feeding observations have been made over intertidal flats and
mangroves. This might be a consistent habitat preference
outside the breeding grounds, as Gull-billed Terns presumed
to be subspecies affinis also appear to be mainly coastal in
places where staging might be expected, e.g. Malaysia
(Medway and Wells 1976), Thailand (Lekagul and Round
1991) and Hong Kong (Carey et al. 2001). Another respect
in which affinis may resemble coast-dwelling migratory
shorebirds of Australasia is their delayed maturity. Our data
indicate they do not moult into a breeding plumage in their
first potential breeding season, instead remaining on the
non-breeding grounds and first migrating north when about
two years old. We do not know the age of first breeding in
macrotarsa, but the low ratio of birds caught with white caps
suggests they attained breeding plumage in their first year.
The ability to breed at an early age would seem a sensible
adaptation in a bird for which breeding is opportunistic.
Subspecies macrotarsa shows some striking behavioural
differences from affinis in north-western Australia, most
obviously in their regular kleptoparasitism of Whimbrels. We
have never seen affinis do this and suspect their small size
would be a disadvantage if they tried to rob such large shore-
birds. Although subspecies macrotarsa is capable of catch-
ing surface-inhabiting prey while in flight, and regularly
does so in inland habitats (e.g. Bourke et al. 1973), on the
intertidal flats of Roebuck Bay, kleptoparasitism is much the
most common feeding method. We have also seen klepto-
parasitism of Whimbrels by macrotarsa on the intertidal flats
of Eighty Mile Beach. It would of interest to know how
widespread such kleptoparasitism is, for Whimbrels occur
on most of the northern Australian coastline where
macrotarsa moves in the austral winter. It is interesting that
on Eighty Mile Beach, where Whimbrels are far less
common than on Roebuck Bay (Rogers, in press;
C. D. T. Minton, personal observation), macrotarsa is also
relatively less common.
There is some regularity to the annual cycle of
macrotarsa: they tend to move to the northern Australian
coast during the austral winter (Blakers et al. 1984; Davies
1988; this study), their breeding plumage is most extensive
in the austral summer (this study; and also observed in south-
eastern Queensland (J. Dening, personal communication))
and most breeding records come from about September to
January (Higgins and Davies 1996). Nevertheless, they are
capable of breeding in the austral winter if conditions are
suitable. Evidence that this applies to the birds in our study
area comes from band recoveries at a breeding colony at
Mandora Marsh in 1999, when Gull-billed Terns began
laying in late May and some were still raising chicks in
October. Other winter breeding events in the region were
observed on Mandora Marsh in May 2004 (J. van de Kam,
personal communication) and on Anna Plains (3 km from
Eighty Mile Beach) in May 2000 (C. D. T. Minton, unpub-
lished observation).
We suggest that the moults and plumages of macrotarsa
reflect adaptations to an opportunistic breeding regime. The
non-breeding plumage of macrotarsa varies a great deal but
Emu 157
the cap is typically mottled, or even fully black as it is in
breeding plumage. The colour of the incoming plumage of
birds can be influenced by their hormonal condition
(Voitkevich 1966), a striking example being shown by the
Black-chested Prinia (Prinia flavicans) in which the basic
plumage is usually drab, but becomes brightly coloured (like
alternate plumage) if breeding occurs during the pre-basic
moult (Herremans 1999). In bird species in which the photo-
periodic controls of breeding activity can be overridden by
other information to permit opportunistic breeding, there is
evidence that relatively high levels of reproductive or
hypothalmic hormones are maintained so gonadal regression
is more easily reversed by stimuli other than photoperiod
(Astheimer and Buttemer 2002). Such cues might include
nutritional status or social interactions (Astheimer and
Buttemer 2002). Accordingly, macrotarsa could be a bird in
which there are higher resting levels of reproductive hor-
mones than in migratory affinis, with these hormones influ-
encing the colour of the cap in pre-basic moult. Some
evidence that the plumage colour of macrotarsa is relatively
sensitive to cues other than photoperiod comes from 15 cap-
tives at Perth Zoo, which were kept in the same enclosure and
remained black-capped for several years (R. E. Johnstone,
personal communication).
Subspecies macrotarsa and affinis also differed in their
primary moult strategies. Terns rarely moult their primaries
while breeding, presumably because of the conflicting
energy costs (Ashmole 1968; Cramp 1985; Higgins and
Davies 1996). Subspecies affinis in Australia typically
showed one active moult locus in the primaries, only sus-
pending primary moult while migrating, and only briefly
showing two moult loci in the primaries when there was
slight overlap between pre-basic and pre-alternate moults of
adults, or between pre-formative and second pre-basic
moults of younger birds. In contrast, macrotarsa showed far
more overlap of pre-basic and pre-alternate primary moults,
with pre-alternate moult beginning when pre-basic moult
had reached ~p5–p6. Moult of one or both of these moult
waves was often interrupted. As a result of this strategy, the
wings of macrotarsa typically showed a confused mixture of
feathers of different ages. We suggest this is advantageous to
an opportunistic breeding cycle, in that there will never be a
large number of old feathers in the wing at any one time. If
an opportunity to breed arises, it will thus be possible to
interrupt active primary moult for several months without
the risk of large numbers of primaries becoming so worn that
flight is impaired. In this respect, the primary moult strategy
of macrotarsa is reminiscent of several species of tropical
pelagic terns in which breeding can occur at almost any time
of year and primary moult occurs between breeding
attempts. At the most striking extreme, the Christmas Island
(Pacific Ocean) population of White Tern (Gygis alba)
replaces its primaries in a series of outward-moving waves,
each separated by about four feathers from the adjacent
moult waves; after being interrupted while breeding, primary
moult can resume at the point of interruption so the oldest
primaries are the first to be replaced (Ashmole 1968). Such
moult strategies (the continuous stepwise moults of Ashmole
1968) result in such a complex jumble of primaries of dif-
ferent ages that it is difficult to guess how the moult strategy
might have evolved. The primary moults of macrotarsa also
produce a mixture of differently aged primaries and could be
confused with true continuous stepwise moult, but our data
suggest that pre-basic and pre-alternate waves of primary
moult are distinguishable and each appears to begin at p1.
The moult strategy could therefore be one that is suitable for
unpredictable breeding seasons, but is derived from an
ancestral strategy like that seen in affinis.
The substantial differences between macrotarsa and
migratory subspecies of Gull-billed Tern such as affinis
suggest that macrotarsa has been isolated for a long time.
Acomprehensive genetic analysis to examine relationships
of macrotarsa, affinis and the other subspecies of S. nilotica
is overdue. Whatever the true taxonomic situation may be,
macrotarsa and the Asian subspecies of Gull-billed Tern are
very different and there is much to learn about the distri-
bution and behaviour of both. We would encourage future
field guides to provide sufficient detail to enable bird-
watchers to distinguish the subspecies in the field.
Acknowledgments
We thank the Australasian Wader Studies Group for access to
their banding data, Broome Bird Observatory for its role as
a base of banding operations, and both for data on migratory
departures. John Stoate of Anna Plains Station provided
essential access and support during fieldwork on Eighty
Mile Beach. Data on intertidal distributions were collected
by D. I. Rogers as part of a shorebird study supported by
Environment Australia and a post-graduate scholarship from
Charles Sturt University. Capture and banding of Gull-billed
Ter ns in north-western Australia was under licence from the
Western Australian Department of Conservation and Land
Management and numbered metal bands were supplied by
the Australian Bird and Bat Banding Scheme. Adrian Boyle
was instrumental in making Gull-billed Tern catches at
crucial times and provided much helpful discussion. For
access to the museum specimens in their care, we thank Rene
Dekker (Rijksmuseum van Natuurlijke Historie, Leiden),
Rory O’Brien (Museum of Victoria) and Ron Johnstone
(Western Australian Museum). Invaluable discussion and
comment were provided by Ken Rogers, Mike Carter, Chris
Corben, Jill Dening, Dion Hobcroft, David James, Jutta
Leyrer, Ian Nisbet, Peter Pyle, Kees Roselaar, Iain Taylor,
Tony Tree and two anonymous referees.
References
Ashmole, N. P. (1968). Breeding and molt in the White Tern (Gygis
alba) on Christmas Island, Pacific Ocean. Condor 70, 35–55.
Gull-billed Terns in North-western Australia
D. I. Rogers et al.158 Emu
Astheimer, L. B., and Buttemer, W. A. (2002). Changes in latitude,
changes in attitude: some physiological perspectives on Australian
Birds. Emu 102, 19–28. doi:10.1071/MU01031
Barter, M. (2002). ‘Shorebirds of the Yellow Sea: Importance, Threats
and Conservation Status.’ Wetlands International Global Series 9,
International Wader Studies 12, Canberra.
Battley, P. F., Piersma, T., Dietz, M. W., Tang, S., Dekinga, A., and
Hulsman, K. (2000). Empirical evidence for differential organ
reductions during trans-oceanic bird flight. Proceedings of the
Royal Society of London. Series B. Biological Sciences 267,
191–195. doi:10.1098/rspb.2000.0986
Blakers, M., Davies, S. J. J. F., and Reilly, P. N. (1984). ‘The Atlas of
Australian Birds.’(Melbourne University Press: Melbourne.)
Bourke, P. A., Lowe, V. T., and Lowe, T. G. (1973). Notes on the Gull-
billed Tern. Australian Bird Watcher 5, 69–79.
Carey, G. J., Chalmers, M. L., Diskin, D. A., Kennerley, P. R.,
Leader, P. J., Leven, M. R., Lewthwaite, R. W., Melville, D. S.,
Turnbull, M., and Young, L. (2001). ‘The Avifauna of Hong Kong.’
(Hong Kong Bird Watching Society: Hong Kong.)
Coates, B. J. (1985). ‘The Birds of Papua New Guinea. Vol. 1.’ (Dove
Publications: Brisbane, Qld.)
Coates, B. J., Bishop, K. D., and Gardner, D. (1997). ‘A Guide to the
Birds of Wallacea. Sulawesi, The Moluccas and Lesser Sunda
Islands, Indonesia.’(Dove Publications: Brisbane, Qld.)
Cramp, S. (Ed.) (1985). ‘The Birds of the Western Palearctic. Vol. IV.’
(Oxford University Press: Oxford.)
Davies, S. J. J. F. (1988). Nomadism in the Australian Gull-billed Tern.
Proceedings International Ornithology Congress 19, 744–753.
Dement’ev, G. P., and Gladkov, N. A. (1951). ‘The Birds of the Soviet
Union. Vol. 3.’ (English translation, 1969.) (Israel Program for
Scientific Translations: Jerusalem.)
Dickinson, E. C., Kennedy, R. S., and Parkes, K. C. (1991). ‘The Birds
of the Philippines.’ British Ornithologists Union Checklist 12.
(British Ornithologists Union: Tring, UK.)
Étchécopar, R. D., and Hüe, F. (1978). ‘Les Oiseaux de Chine, de
Mongolie et de Corée.’(Editions du Pacifique: Papeete, Tahiti.)
Gochfeld, M., and Burger, J. (1999). Family Sternidae. In ‘Handbook of
the Birds of the World. Vol. 3. Hoatzin to Auks’. (Eds J. del Hoyo,
A. Elliott and J. Sargatal.) pp. 624–627. (Lynx Edicions: Barcelona.)
Herremans, M. (1999). Biannual complete moult in the Black-chested
Prinia Prinia flavescens. Ibis 141, 115–124.
Higgins, P. J., and Davies, S. J. J. F. (1996). ‘Handbook of Australian,
New Zealand and Antarctic Birds. Vol. 3. Snipe to Pigeons.’ (Oxford
University Press: Melbourne.)
Howell, S. N. G., Corben, C., Pyle, P., and Rogers, D. I. (2003). The first
basic problem: a review of molt and plumage homologies. Condor
105, 635–653.
Humphrey, P. S., and Parkes, K. C. (1959). An approach to the study of
moult and plumages. Auk 76, 1–31.
Ingraham, S. (2001). Is there a digital camera in your birding future?
Birding April 2001, 164–171.
Jaensch, R. (1983). Gull-billed Terns in the South-west. Swans 13(3),
23–26.
Johnstone, R. E. (1977). An Asian Gull-billed Tern in Western
Australia. Western Australian Naturalist 14, 26–27.
Johnstone, R. E., and Storr, G. M. (1998). ‘Handbook of Western
Australian Birds. Vol. 1 – Non-passerines (Emu to Dollarbird).’
(Western Australian Museum: Perth.)
Lane, B., and Jessop, A. (1985). Tracking of migrant waders in north-
western Australia using meteorological radar. Stilt 6, 17–28.
Lekagul, B., and Round, P. D. (1991). ‘A Guide to the Birds of
Thailand.’(Saha Karn Bhaet: Bangkok.)
Marchant, S., and Higgins, P. J. (1990). ‘Handbook of Australian, New
Zealand and Antarctic Birds. Vol. 1A. Ratites to Petrels.’ (Oxford
University Press: Melbourne.)
McKean, J. L. (1981). The status of gulls and terns (Laridae) in the
Darwin area, N.T. 1974 to 1980. Australasian Seabird Group
Newsletter 15, 11–16.
Medway, Lord, and Wells, D.R. (1976). ‘The Birds of the Malay
Peninsula. Vol. 5.’ (Witherby: London.)
Mees, G. F. (1982). Birds from the lowlands of southern New Guinea
(Merauke and Koemba). Zoologische Verhandelingen 191, 1–188.
Meyer de Schauensee, R. (1984). ‘The Birds of China.’ (Oxford
University Press: Oxford.)
Olsen, K. M., and Larsson, H. (1995). ‘Terns of Europe and North
America.’(Christopher Helm: London.)
Peters, J. L. (1934). ‘Check-list of Birds of the World. Vol. 2.’ (Museum
of Comparitive Zoology: Cambridge, MA.)
Piechocki, R. (1968). Zur Avifauna der Mongolei, I. Mitteilungen aus
dem Zoologischen Museum in Berlin 44, 149–292.
Piersma, T., Zwarts, L., and Bruggemann, J. H. (1990). Behavioural
aspects of the departure of waders before long-distance flights:
flocking, vocalizations, flight paths and diurnal timing. Ardea 78,
157–184.
Rogers, K. G. (1995). SHEBA: computer programs for sexing birds on
measurements using univariate data. Corella 19, 25–34.
Rogers, D. I. (1999). What determines shorebird feeding distribution in
Roebuck Bay? In ‘Intertidal Sediments and Benthic Animals of
Roebuck Bay, Western Australia’. (Eds M. Pepping, T. Piersma,
G. Pearson and M. Lavaleye.) pp. 145–174. NIOZ-report series
1999-3. (Texel, The Netherlands.)
Rogers, D. I. (In press). Distribution of shorebirds along Eighty-mile
Beach. In ‘The Long Mud. Benthos and Shorebirds of the Foreshore
of Eighty-mile Beach, Western Australia’. (Eds G. Pearson,
T. Piersma, R. Hickey and M. Lavaleye.) (Royal Netherlands
Institute for Sea Research; Western Australian Department of
Conservation and Land Management, Environs Kimberley: Perth;
and Central Washington University: Ellensburg, WA.)
Rogers, D. I. (2004). Sorting migrant Gull-billed Terns from residents.
Wingspan 14(4), 22–25.
Schodde, R. (1991). The Asian Gull-billed Tern Sterna nilotica affinis
in Australia. Bulletin of the British Ornithologists’ Club 111,
215–217.
Stewart, D. (1999). ‘Bird Calls of the Broome Region.’ (Nature Sound:
Mullumbimby, NSW.)
Vaurie, C. (1965). ‘The Birds of the Palearctic Fauna. Vol. 2. Non-
passeriformes.’(Witherby: London.)
Voitkevich, A. A. (1966). ‘The Feathers and Plumage of Birds.’
(Sidgwick and Jackson: London.)
White, C. M. N., and Bruce, M. D. (1986). ‘The Birds of Wallacea.’
British Ornithologists’ Union Checklist 7. (British Ornithologists’
Union: London.)
Manuscript received 10 October 2004, accepted 4 April 2005
http://www.publish.csiro.au/journals/emu
