Moult and plumage

Abstract

Moult is one of the three main energy‐demanding events in the yearly cycles of birds, and in most species occurs at a different time from breeding and migration. The sequence in which these events occur varies according to the ecological circumstances in which particular populations live, and in general moult is more variable in timing than other events. Some migratory birds moult in their breeding areas after nesting is over; others moult at a staging area on migration; while others moult in winter quarters. Yet others show a split moult, replacing part of their plumage in one place and part in another, moult being arrested during the intervening migration. Different variants in these patterns occur in different populations of the same species. Some types of birds overlap breeding and moult, and some also overlap moult and migration, especially body moult which can occur without reducing flight efficiency. Most of our knowledge of moult is based on museum skins and, now that appropriate statistical models are available to analyse it, moult is increasingly being studied in free‐living birds, the main parameters of ecological interest being its timing and duration. Recent models for analysing moult data, and some of the pitfalls, are discussed briefly in this paper. More field study of moult is needed, particularly because of its relevance to annual cycle and population processes, and hence to its potential application in conservation‐based research. Ringers are the only people at present able to fill this gap in knowledge.

Full text

© 2009 British Trust for Ornithology
Ringing & Migration (2009) 24, 220–226
Moult and plumage
IAN NEWTON*
Centre for Ecology and Hydrology, MacLean Building, Benson Lane, Crowmarsh Gifford,
Wallingford, Oxfordshire OX10 8BB
Moult is one of the three main energy-demanding events in the yearly cycles of birds, and in most species
occurs at a different time from breeding and migration. The sequence in which these events occur varies
according to the ecological circumstances in which particular populations live, and in general moult is more
variable in timing than other events. Some migratory birds moult in their breeding areas after nesting is
over; others moult at a staging area on migration; while others moult in winter quarters. Yet others show a
split moult, replacing part of their plumage in one place and part in another, moult being arrested during
the intervening migration. Different variants in these patterns occur in different populations of the same
species. Some types of birds overlap breeding and moult, and some also overlap moult and migration,
especially body moult which can occur without reducing flight efficiency. Most of our knowledge of moult
is based on museum skins and, now that appropriate statistical models are available to analyse it, moult is
increasingly being studied in free-living birds, the main parameters of ecological interest being its timing
and duration. Recent models for analysing moult data, and some of the pitfalls, are discussed briefly in
this paper. More field study of moult is needed, particularly because of its relevance to annual cycle and
population processes, and hence to its potential application in conservation-based research. Ringers are
the only people at present able to fill this gap in knowledge.
* Email: ine@ceh.ac.uk
My aim in this paper is to describe how moult (the period
of plumage renewal) fits within the annual calendar of
different bird species, how it can be studied during ringing
operations, and analysed using appropriate statistical
methodology to estimate its timing and duration. Although
moult is one of the three major events in the annual cycles
of birds, it has been much less studied than breeding and
migration (for reviews see Stresemann & Stresemann 1966,
Jenni & Winkler 1994, Kjellén 1994, Newton 2008). This
is partly because of a general lack of appreciation of its
relevance to the annual cycle and population processes,
and its potential role in conservation-driven research.
Feathers are special lightweight structures consisting of
a tough, inert protein called keratin. They form one of
the defining features of birds, whether the soft insulating
feathers that cover the body surface or the stiffer flight
and tail feathers that provide the aerofoils which permit
flight. Once they are formed, feathers become attached
dead structures in which damaged parts cannot be repaired.
They deteriorate mainly through the action of physical
wear, sunlight and feather mites, and must therefore be
renewed periodically. During a moult, feathers are generally
replaced sequentially, in predetermined order, so that
body insulation and (in most birds) flight are maintained
throughout. Each feather is shed as a new one begins to
grow below, and each has a characteristic growth curve,
taking its own set period to reach full length. Within
species, equivalent feathers in different individuals take
about the same time to grow, so that individual variation
in moult duration is due much more to variations in the
intervals between the shedding of successive feathers than
to variation in the growth rates of their replacements (eg
Newton 1967, 1969, Serra 2000). Many birds moult only
once a year, others two or more times. Moults may be
complete, involving body, wing and tail feathers, or partial,
involving body feathers alone (and sometimes a few flight
feathers).
MOULT AS A COMPONENT OF THE ANNUAL
CYCLE
The processes of breeding, moult and migration all require
extra food above the needs of daily maintenance and in
many birds occur mainly at different seasons. Moult cannot,
therefore, be considered in isolation from other events in
the annual cycle, and to a large extent the three main events
should be viewed as an integrated whole. In addition, many
birds show a quiescent period in winter, during which
they are not breeding, moulting or migrating. Outwardly,
they seem to be doing little except eating and surviving,
but inwardly they may be undergoing some physiological
change, such as growing gonads in preparation for breeding
at a later date. Not all species show this quiescent stage,

© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
Moult and plumage 221
Figure 1. Eight common sequences of annual cycle events described
among European migratory birds. The term ‘quiescence’ denotes a
period when the bird is not breeding, moulting or migrating.
Pre-breeding migration – breeding – moult – post-breeding
migration. Examples: Chaffinch Fringilla coelebs, Common
Redpoll Carduelis flammea, Thrush Nightingale Luscinia luscinia,
Fieldfare Turdus pilaris, Jack Snipe Lymnocryptes minimus.
Pre-breeding migration – breeding – post-breeding migration
part 1 – moult – post-breeding migration part 2. Examples:
Great Reed Warbler Acrocephalus arundinaceus, River Warbler
Locustella fluviatilis, Northern Lapwing Vanellus vanellus, some
populations of Green Sandpiper Tringa ochropus.
Pre-breeding migration – breeding – post-breeding migration
– moult. Examples: Common Rosefinch Carpodacus erythrinus,
Barn Swallow Hirundo rustica, Common Swift Apus apus,
Whimbrel Numenius phaeopus, Harlequin Duck Histrionicus
histrionicus, together with some shearwaters, terns and skuas
that breed and winter in opposite hemispheres. In some
of these species, the moult is prolonged and the quiescent
period short or non-existent (as in Barn Swallows wintering
in South Africa).
Pre-breeding migration – breeding – moult part 1 – post-breeding
migration – moult part 2. Examples: Barred Warbler Sylvia
nisoria, Alpine Swift Apus melba, Scops Owl Otus scops,
1.
2.
3.
4.
however, as some pass without obvious break from one
major process to another (Newton 2008).
At least eight different sequences of events occur
commonly among different European bird populations
and others less commonly, depending on the particular
ecological circumstances in which each population lives
(Fig 1). In general, residents and short-distance migrants
moult in summer after breeding (residents more slowly: Fig
1, sequence 1). Long-distance migrants moult either in late
summer in the breeding area (as in sequence 1), in autumn
1
Post-breeding
migration
Breeding
Pre-breeding
migration
Quiescence
Moult
2
Post-breeding
migration, phase 2
Breeding
Pre-breeding
migration
Quiescence
Moult
Post-breeding
migration, phase 1
3
Post-breeding
migration
Breeding
Pre-breeding
migration
MoultQuiescence
4
Post-breeding
migration
Moult,
phase 1
Moult,
phase 2
Breeding
Pre-breeding
migration
Quiescence
5
Post-breeding
migration, phase 1
Post-breeding
migration, phase 2
Moult,
phase 1
Moult, phase 2
Breeding
Pre-breeding
migration and
moult
Quiescence
Post-breeding
migration, phase 1
6
Moult, phase 1
Moult, phase 2
Breeding
Quiescence
Post-breeding
migration, phase 2
Pre-breeding
migration and moult
7
Breeding
Pre-breeding
migration
Quiescence
Post-breeding
migration
Post-nuptial
moult
Pre-nuptial
moult
Breeding
8
Pre-breeding
migration
Quiescence
Post-breeding
migration
Post-nuptial
moult
Pre-nuptial
moult
at a migratory staging area (sequence 2) or in winter quarters
(sequence 3) (Bensch et al 1991, Jenni & Winkler 1994,
Newton 2008). Moulting in wintering areas is widespread
among northern-hemisphere species which spend their
non-breeding period in the southern hemisphere, where
the seasons are reversed (the southern summer coinciding
with the northern winter). Less constrained by time, they
also spread the moult over a longer period.
In many other migratory species, the moult occurs partly
in one area and partly in another, separated by migration.
Bee-eater Merops apiaster, Red-necked Nightjar Caprimulgus
ruficollis, Collared Pratincole Glareola pratincola, Marsh
Sandpiper Tringa stagnatilis, juveniles of Common Quail
Coturnix coturnix.
Pre-breeding migration – breeding – moult part 1 – post-breeding
migration part 1 – moult part 2 – post-breeding migration part
2. Examples: Kentish Plover Charadrius alexandrinus, Spotted
Redshank Tringa erythropus, some individuals of Curlew
Sandpiper Calidris ferruginea.
Pre-breeding migration – breeding – post-breeding migration
part 1 – moult part 1 – post-breeding migration part 2 – moult
part 2. Examples: Wilson’s Phalarope Phalaropus tricolor,
Spotted Sandpiper Actitis macularia, and other populations of
shorebirds.
Pre-breeding migration – breeding – first (post-nuptial) moult
– post-breeding migration – second (pre-nuptial) moult. Example
with two complete moults per year: Willow Warbler Phylloscopus
trochilus; examples with one complete and one partial moult
per year: Melodious Warbler Hippolais polyglotta and many
shorebirds.
Pre-breeding migration – breeding – post-breeding migration
– post-nuptial moult – pre-nuptial moult. Examples: Lanceolated
Warbler Locustella lanceolata and some other Locustella
warblers, Curlew Sandpiper Calidris ferruginea.
5.
6.
7.
8.

222 I. Newton
© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
The moult can be split between the breeding area and
wintering area (sequence 4), between the breeding area
and a staging area (sequence 5), or between a staging and
wintering area (sequence 6). The moult normally stops
during migration, so that the bird can fly with a full set
of flight feathers, some new and others old. The bird
resumes the second part of moult wherever it left off in
the first part (with few exceptions). In the last two of these
patterns (sequences 5 and 6), a split moult is associated
with a split migration. In other (mostly large) species,
split moults are associated with breeding (as moult stops
temporarily during chick feeding), or with periods of winter
food shortage. Comparing these various patterns among
species, moult is much more variable in timing than is
breeding or migration, probably because moult scheduling
is less crucial.
The story does not end here, for while some migratory
species have a single split moult, replacing their feathers
once but in two bouts, other species have two separate
moults, replacing the same feathers twice in one year. One
moult occurs either before or after autumn migration (the
post-nuptial moult), and the other before and during spring
migration (the pre-nuptial moult) (sequences 7 and 8). In
most twice-yearly moulting species, the autumn moult is
complete and the spring moult is partial, involving body
feathers only (and sometimes a few tertial, secondary or
tail feathers). However, in a small proportion of species
that moult twice each year, such as the Willow Warbler
Phylloscopus trochilus, both moults are complete, involving
the replacement of both body and wing feathers. In some
species with two moults per year, both plumages look the
same (as in Willow Warbler), but in other species pre-
nuptial moult gives rise to a special breeding plumage,
more brightly coloured than the drab winter garb. Some
species, such as the Linnet Carduelis cannabina and
Brambling Fringilla montifringilla, acquire their breeding
plumage, not by a pre-nuptial moult, but by abrasion, in
which dull feather tips wear off to expose colour below. Pre-
nuptial moults of the body feathers occur in many species
of waders, and usually overlap with spring migration. In
diving and dabbling ducks, the pre-nuptial moult (mainly
body feathers) follows a few weeks after the post-nuptial
moult (complete). In consequence, drakes are in dull
‘eclipse’ plumage (equivalent to winter plumage) for only
a few weeks each year and in bright breeding plumage for
most of the year (Cramp & Simmons 1977, Bluhm 1988).
In association with this, many species of ducks form pairs
while in winter quarters, whereas most other birds pair up
in breeding areas.
These various generalisations apply mainly to small or
medium-sized birds, in which moult occurs as a distinct
event in the annual cycle, typically lasting two to three
months (Fig 1). In many species, moult, breeding and
migration each occupy short enough periods that they can
all be fitted within a year without overlap, and often with a
quiescent period as well. In some large birds, however, such
as vultures and albatrosses, breeding cycles and moult take
so long that they cannot both be fitted within a calendar
year without overlapping, and in some such species moult
may also overlap with migration, especially body moult
which does not reduce flight efficiency (Stresemann &
Stresemann 1966).
In most raptors, moult begins during incubation (earlier
in females than males) and overlaps with most of the
breeding cycle, although it may be arrested during chick
rearing (as in the Sparrowhawk Accipiter nisus, Newton
& Marquiss 1982). Smaller raptor species can normally
finish their moult before the post-breeding migration,
but larger ones, which take longer to grow their feathers,
arrest moult during migration, and continue after reaching
winter quarters (as in the Osprey Pandion haliaetus and
Honey Buzzard Pernis apivorus). In some of the largest flying
birds, such as vultures, condors, storks and albatrosses,
each moult cycle lasts more than a year, but again may
be interrupted during difficult periods, such as chick-
rearing. Otherwise such birds appear to moult more or less
continuously, and may have two or more moult waves in
the primary and secondary flight feathers at once (so-called
serial moult). In addition, some large aquatic birds, such
as waterfowl and grebes, circumvent the problem of slow
feather growth in a different way, by moulting all their flight
feathers simultaneously (becoming temporarily flightless).
The whole feather series is then replaced within the time
taken to grow the longest primary (about four weeks in
ducks, six weeks in geese).
Birds clearly show great variation between species in the
sequence of events through the year, their duration and
extent of overlap. Moreover, unlike a successful breeding
attempt, moult and migration can be stopped while the
bird does something else. This facility adds yet more
variation to the range of annual schedules found among
birds, fitting the various patterns in food availability and
risk to which different migratory populations are exposed
during the year. This variation in annual schedules is shown
mostly in comparisons between species, but also to some
extent between different geographical populations of the
same species. For example, with increasing latitude, the
migrations of many species lengthen, and take up more
of the year, while the periods devoted to breeding and
moult decline in association with the decreasing length
of the favourable season. In some species, populations at
lower latitudes moult in breeding areas, whereas those from
higher latitudes postpone their moult for winter quarters.
Thus, Barn Swallows Hirundo rustica in the southernmost

Moult and plumage 223
© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
breeding areas, which are resident or short-distance
migrants, moult during June–August after breeding;
whereas those in the most northerly breeding areas begin
moulting in September–October, after they have reached
their distant wintering areas. At intermediate latitudes
(including Britain and Ireland), varying proportions of
individuals show a split moult, starting in breeding areas,
arresting during migration, and resuming in winter quarters
(Cramp 1988). Likewise, most European populations of
Ringed Plovers Charadrius hiaticula moult rapidly in their
breeding areas in August–September, before migrating
short distances within Europe, whereas arctic-nesting
birds leave their nesting areas after breeding, and postpone
their moult until November–March after reaching their
wintering areas in southern Africa (Stresemann &
Stresemann 1966). Other geographical variants in the
timing and duration of moult occur in many other wader
species, mainly in association with the latitudes at which
they breed and winter (see Cramp & Simmons 1983,
Serra 1998, Underhill 2003). Some species also show sex
differences in the timing of moult and migration, according
to their different parental roles (Newton 2008). Otherwise,
individual variations in the start dates of moult in the
same population relate chiefly to variations in the dates
they finish their preceding activities. Among populations
which moult in their breeding areas, adults that continue
breeding later in the year than others start their moult later,
and young raised late in the year start moulting later than
earlier-hatched young (eg Newton 1966, 1999, Newton &
Rothery 2005, Flinks et al 2008). It is also common for late-
nesting adults to start moulting while they still have young
in the nest, and to replace their feathers more rapidly or
less completely than earlier-moulting individuals. Similarly,
late-fledged juveniles start moulting at an earlier age than
early-hatched ones, thereby lessening the delay in their
migration (Jenni & Winkler 1994, Newton 2008).
THE RECORDING OF MOULT
Most of our knowledge of moult timing, of the kind
mentioned above, is based on generalisations made from
museum skins. Until the 1960s we had no method of
systematically recording the state and progress of moult
in a way that would enable its timing and duration in a
population to be estimated accurately. However, it was clear
that, in many birds, moult of the primary flight feathers
spanned the whole (or almost the whole) moult period; they
were shed in sequence through the series, so that, for most
of the moult, several primaries were in growth at once, at
different stages. A recording system was therefore devised
in which each primary in one wing was given a score,
according to its stage of growth: old feathers were scored as
0, new ones as 5, and growing ones as 1–4. Adding together
the scores of the different primaries produced a single score
reflecting the stage of moult in the individual concerned
(Ashmole 1962, Evans 1966, Newton 1966). Species with
nine large primaries in each wing scored a maximum of 45
per wing, and species with ten large primaries scored 50.
On this system, the intensity of moult could also be assessed
from the number of flight feathers growing simultaneously
(or from the ‘residual raggedness score’: Bensch & Grahn
1993). Because wings normally moult in step with one
another, it is not necessary to record both. This was the
method used in the British Trust for Ornithology’s moult
recording scheme started in the 1960s.
Regression methods
Plotting the scores of different birds against date enabled
the mean start date and mean rate (or duration) to be
estimated. Regression methods were used in early studies
(Evans 1966, Newton 1966, Ginn & Melville 1983), but
were unsatisfactory because the regression line tends to
run diagonally across the long axis of the parallelogram
enclosing the scatter of points. It effectively gave the
start and end dates for the population as a whole, rather
than for the average bird, thus underestimating the mean
start date and overestimating the mean duration (Pimm
1976, Summers et al 1983). More realistic estimates of
start dates and duration were obtained by calculating the
regression of date on score. Such estimates tend to give
the closest fit to estimates of moult rate obtained from
examining individual birds more than once during moult,
but this approach does not meet some of the assumptions
of standard regression analysis (Underhill & Zucchini
1988). It tends to underestimate mean start date and
overestimate mean duration, especially in populations
in which individuals show wide variation in start dates.
Moreover, regression methods were based only on birds
in active moult, ignoring birds which had not started or
had finished, and thus discarded some potentially useful
information.
Underhill–Zucchini models for avian moult
In response to these problems, Underhill & Zucchini
(1988) proposed a model for avian moult data which made
use of non-moulting as well as moulting birds, fitting the
model using the method of maximum likelihood. They
considered three data types: Type 1, each bird placed in one
of three categories – moult not started, in moult and moult
finished; Type 2, each bird classified as in Type 1, except
that moulting birds were given a moult score to reflect the
stage of moult; and Type 3, in which each bird in moult
was given a score, but non-moulting birds were ignored
(as in the early regression methods). The underlying
assumptions were that: (a) birds caught on each day were

224 I. Newton
© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
a random sample from the relevant population, (b) the
times of onset of moult followed a specified distribution,
such as the Normal Distribution, and (c) for Type 2 and
Type 3 data the moult score increased linearly with time,
and consistently between individuals.
In an analysis of Bullfinch Pyrrhula pyrrhula moult, the
effects on estimated moult parameters of deviations from
these assumptions were explored using simulated data
(Newton & Rothery 2000). It emerged that only slight
deviation from linearity in Type 3 data had substantial
effects on estimates. The most reliable estimates for this
species were obtained by using Type 1 data, ignoring scores.
Among birds in general, the rate of increase in moult
score is seldom expected to be linear throughout moult,
partly because near the start and end the bird has only one
primary per wing in growth, whereas for the rest of moult
it has up to several at once. The progress of moult would
therefore be expected to follow an S-shaped curve, slower
at the start and end than in the middle (as confirmed in
studies of moult based on captive birds: Newton 1967,
Dawson & Newton 2004).
The problem of non-linearity is accentuated in some
birds, such as waders, in which different primary feathers
vary greatly in length, with long outer ones taking more
than twice as long to grow as short inner ones (Summers et
al 1983). One method devised to correct for this variation
is to weigh each of the primary feathers (obtained from
dead birds), and then correct the scores of living birds to
an appropriate weight of new feather material produced
(Summers et al 1983, Underhill & Joubert 1995).
The total score of a bird obtained visually can then be
converted to a feather mass score, reflecting the percentage
weight of new feather material produced. Such ‘feather
mass scores’ (FMS), or ‘percentage feather mass grown’
(PFMG: Summers et al 1983) generally give a more linear
relationship with date than do the original scores based
on feather lengths (as also confirmed in five species of
passerines: Dawson & Newton 2004). They make a bigger
difference to estimates of moult start dates and durations
for waders, whose feathers vary more in length than those
of most passerines. Some researchers have additionally
allowed for the fact that feathers vary in structure along
their length, calculating a separate weight for each part of
each feather (Redfern 1998). Other potential approaches
to allowing for the non-linearity in the moult score or
PFMG are (a) to transform the data in some way, and (b)
to extend the Underhill–Zucchini model to allow for non-
linear increase in moult scores.
Testing the validity of different methods depends
critically on estimates of the rate of increase in moult
score obtained from individuals caught more than once
during moult. These individual figures give direct measures
of moult rate which are immune to the biases that affect
overall estimates of moult rate obtained from scattergrams
of moult score against date. However, in most studies such
data are limited because few individuals are retrapped.
Comparing Type 1, 2 and 3 values against individual values
for three species (Starling Sturnus vulgaris, Bullfinch and
Sanderling Calidris alba) showed that estimates of start and
end dates (and hence durations) based on Type 1 data were
either better or as good as those obtained using Types 2
and 3 data (Rothery & Newton 2002). Type 1 data are the
easiest to collect because they do not need the allocation of
a moult score or its conversion to mass. The Type 1 method
could also be used for partial moults (including post-
juvenile moult) in which flight feathers are not replaced. It
is easy to implement using the standard binary regression
models that are available in most statistical packages (such
as Minitab Release 12).
Methods that depend on recording non-moulting as well
as moulting birds work well for species that are resident
year-round in the same area, for samples can then be drawn
from the entire population throughout. But sometimes
birds start moulting soon after they have moved into an
area, and stay on later, in which case non-moulting birds
are under-represented in the initial period before all have
arrived. One way round this problem is to restrict analysis
to moulting and post-moult birds, omitting any pre-moult
birds (the Type 4 method of Underhill et al 1990). In
other situations, birds leave on migration after completing
moult, so non-moulting birds are under-represented in the
later period, once the first birds start to leave. In this case
Figure 2. Numerical primary scores of Goldfinches plotted in relation
to date, with the line fitted using the Underhill–Zucchini procedure,
Type 5, which excluded birds that had finished moult (• - males; -
females). Mean start date was estimated as 25 July, with 95% of
the birds starting within the 31-day period, 9 July – 9 August, and
finishing within the period 22 September – 24 October. The mean
duration was 76 days (se = ±4 days). No significant sex differences
emerged. From Newton & Rothery 2009.
∧

Moult and plumage 225
© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
analysis is best restricted to pre-moulting and moulting
birds (the Type 5 method of Underhill et al 1990). This
last procedure was used to estimate moult parameters in
Stonechats Saxicola torquatus and Goldfinches Carduelis
carduelis, because it was suspected that many birds left the
study area on completing their moult (Fig 2; Flinks et al
2008, Newton & Rothery 2009).
The Underhill–Zucchini procedure has provided a
means to obtain reliable estimates of moult timing and
rate (and their standard deviations) from only two or three
categories of birds (pre-moult, in moult, post-moult). It has
been used in at least 30 studies so far. In some species in
which it was tested, the Type 1 method gave estimates at
least as reliable as any measures which included the use
of moult scores. However, further testing is required on a
wider range of species, and including separate analyses for
juveniles undergoing body moult alone. Almost certainly,
different models will prove the most appropriate for
different species. This is mainly because species vary in the
extent to which they deviate from each of the assumptions
underlying the models, and different models vary in their
sensitivity to each assumption.
CONCLUDING REMARKS
For most European birds we still lack detailed studies of
moult in which the data are adequate for statistical analysis.
Yet moult is one of the easiest things for ringers to record,
especially if they ignore moult scores, and instead place
each bird in one of three categories: pre-moult, in moult,
and post-moult. Providing the population can be sampled
appropriately throughout the moult period, such data
should be sufficient to provide reliable estimates of moult
start dates and durations for the population (with separate
estimates for juveniles changing only their body feathers).
Only with sustained study over the years will we reach a
position in which annual variation and time trends in
these parameters can be examined as a matter of routine.
We now have detailed information on the annual
breeding of many bird species in Britain and Ireland, but
almost nothing on the equivalent annual variation in
moult. What are we missing? How much does the spread
of moult within a population, its timing and duration
vary between years, habitats or regions? And how does
this variation relate to breeding and other events in the
annual cycle, or to environmental variables, including
climate? If we are to venture into this enticing field of
avian ecology, it is bird ringers who must take the lead,
for only they are handling sufficient numbers of live birds
year after year in the same places. Moreover, because of
its close interconnection with breeding and migration,
moult can be used to throw light on aspects of bird ecology
otherwise hard to study. For example, in some multi-
brooded passerine species, the amount of late breeding
can vary greatly from year to year. In species which moult
immediately after breeding (like most British passerines),
this annual variation can be assessed much more easily
by obtaining estimates of moult start dates in different
years than by the more laborious procedure of late-season
nest-searching. Among Bullfinches studied over five years
near Oxford, the proportion of adults which started
moult after 20 August (implying successful breeding from
eggs laid after mid July) varied between 7% and 68% in
different years. End-of-season young-to-adult ratios varied
twofold between years, according to the amount of late
nesting (Newton 1999). This story would not have emerged
without a detailed and fascinating, but undemanding,
study of moult.
ACKNOWLEDGEMENTS
I am grateful to Lukas Jenni, Chris Redfern and Peter Rothery,
and an anonymous referee for helpful comments on the
manuscript.
REFERENCES
Ashmole. N.P. (1962) The Black Noddy Anous tenuirostris on Ascension
Island. Part I. General Biology. Ibis 103, 235–273.
Bensch, S. & Grahn, M. (1993) A new method for estimating individual
speed of molt. Condor 95, 305–315.
Bensch, S., Hasselquist, D., Hedenström, A. & Ottosson, U.
(1991) Rapid moult among palaearctic passerines in West Africa – an
adaptation to the oncoming dry season? Ibis 133, 47–52.
Bluhm, C. (1988) Temporal patterns of pair formation and reproduction
in annual cycles and associated endocrinology in waterfowl. Current
Ornithology 5, 123–185.
Cramp, S. (1988) Handbook of the Birds of Europe, the Middle East and
North Africa. Volume 5. Oxford University Press, Oxford.
Cramp, S. & Simmons, K.E.L. (1977) Handbook of the Birds of
Europe, the Middle East and North Africa. Volume 1. Oxford University
Press, Oxford.
Cramp, S. & Simmons, K.E.L. (1983) Handbook of the Birds of
Europe, the Middle East and North Africa. Volume 3. Oxford University
Press, Oxford.
Dawson, A. & Newton, I. (2004) Use and validation of a molt score
index corrected for primary-feather mass. Auk 121, 372–379.
Evans, P.R. (1966) Autumn movements, moult and measurements of the
Lesser Redpoll Carduelis flammea cabaret. Ibis 108, 183–216.
Flinks, H., Helm, B. & Rothery, P. (2008) Plasticity of moult and
breeding schedules in migratory European Stonechats Saxicola
rubicola. Ibis 150, 687–697.
Ginn, H.B. & Melville, D.S. (1983) Moult in birds. BTO Guide 19.
British Trust for Ornithology, Tring.
Jenni, L. & Winkler, R. (1994) Moult and ageing of European
passerines. Academic Press, London.
Kjellén, N. (1994) Moult in relation to migration in birds – a review.
Ornis Svecica 4, 1–24.

226 I. Newton
© 2009 British Trust for Ornithology, Ringing & Migration, 24, 220–226
Serra, L. (1998) The adaptation of primary moult to migration and
wintering in the Grey Plover (Pluvialis squatarola), a preliminary
outlook. Biologia e Conservazione della Fauna 102, 123–127.
Serra, L. (2000) How do Palaearctic Grey Plovers adapt primary moult
to time constraints? An overview across four continents. Wader Study
Group Bulletin 93,11–12.
Stresemann, E. & Stresemann, V. (1966) Die Mauser der Vögel.
Journal für Ornithologie 107, 3–448.
Summers, R.W., Swann, R.L. & Nicoll, M. (1983) The effects of
methods on estimates of primary moult duration in the Redshank Tringa
totanus. Bird Study 30, 149–156.
Underhill, L.G. (2003) Within ten feathers: primary moult strategies
of migratory waders (Charadrii). In Avian Migration (eds Berthold,
P., Gwinner, E. & Sonnenschein, E.), pp 187–197. Springer-Verlag,
Berlin.
Underhill, L. & Joubert, A. (1995) Relative masses of primary feathers.
Ringing & Migration 16, 109–116.
Underhill, L.G. & Summers, R.W. (1993) Relative masses of primary
feathers in waders. Wader Study Group Bulletin 71, 29–31.
Underhill, L.G. & Zucchini, W. (1988) A model for avian primary
moult. Ibis 130, 358–372.
Underhill, L.G., Zucchini, W. & Summers, R.W. (1990) A model
for avian primary moult data types based on migration strategies and
an example using the redshank Tringa totanus. Ibis 132, 118–123.
Newton, I. (1966) The moult of the Bullfinch Pyrrhula pyrrhula. Ibis
108, 41–67.
Newton, I. (1967) Feather growth and moult in some captive finches.
Bird Study 14, 10–24.
Newton, I. (1969) Moults and weights of captive Redpolls. Journal für
Ornithologie 110, 53–61.
Newton, I. (1999) An alternative approach to the measurement of
seasonal trends in bird breeding success: a case study of the Bullfinch
Pyrrhula pyrrhula. Journal of Animal Ecology 68, 698–707.
Newton, I. (2008) The migration ecology of birds. Academic Press,
London.
Newton, I. & Marquiss, M. (1982) Moult in the Sparrowhawk. Ardea
70, 163–172.
Newton, I. & Rothery, P. (2000) Timing and duration of moult in
the Bullfinch Pyrrhula pyrrhula: an appraisal of different analytical
procedures. Ibis 142, 65–74.
Newton, I. & Rothery, P. (2005) The timing, duration and pattern of
moult and its relationship to breeding in a population of the European
Greenfinch Carduelis chloris. Ibis 147, 667–679.
Newton, I. & Rothery, P. (2009) Timing and duration of moult in adult
European Goldfinches. Bird Study 56, 282–288.
Pimm, S. (1976) Estimation of the duration of bird molt. Condor 78,
550.
Redfern, C.P.F. (1998) The analysis of primary moult using feather mass.
Ringing & Migration 19, 39–40.
Rothery, P. & Newton, I. (2002) A simple method for estimating
timing and duration of avian primary moult using field data. Ibis
144, 526–528.