Primary Moult Research Articles

Age differences in primary moult of the Black-headed Gull Chroicocephalus ridibundus

Flight feather moult is an energetically expensive stage of the annual cycle of birds. Its timing is adjusted to other important time- and energy-demanding activities, including migration. In the vast majority of migratory birds, primary moult occurs before or after migration or moulting is suspended during migration. The Black-headed Gull is an exception: it starts moulting flight feathers at the onset of the autumn migration. The course of primary moult in this species is still poorly understood due to the difficulty in obtaining sufficient data from the entire period of flight feather replacement. In this paper, we employed digital photos taken in the field to estimate the start, variation in the start date and duration of primary moult in adult and immature Black-headed Gulls, by using the Underhill–Zucchini likelihood moult model. On average, immatures started their moult on 7 June, 25 days earlier than adults, but their moult lasted 5 days longer. Consequently, there was a 20-day difference in the end of moult – that is, on average it occurred on 16 September and 6 October in immature and adult gulls, respectively. Adults showed greater variability in the mean primary moult start date than immatures, as the beginning of flight feather replacement may depend on breeding success (earlier after breeding failure) and on breeding phenology. The overlap of primary moult and migration produces a trade-off in resource allocation between those two processes, which leads to a low number of simultaneously growing primaries and a decrease in the growth rate of the two external, heaviest flight feathers. Existing reports of a later completion of primary moult in the first half of November in this species are based on the faulty assumption that the outermost primary is always the longest, which is not the case in 34% of individuals.

Primary Moult of Eurasian Spoonbills Platalea l. leucorodia in the Wadden Sea in Relation to Age, Breeding and Migration

Primary Moult of Eurasian Spoonbills Platalea l. leucorodia in the Wadden Sea in Relation to Age, Breeding and Migration

Global review of quantitative studies of primary moult of birds using the Underhill-Zucchini moult model

Knowledge about the timing of moult and how it fits into the annual cycle, along with breeding and, in some cases migration, is an important component of understanding the life history of birds. In comparison with breeding and migration, there is a lack of knowledge about the moult of birds. Primary moult can be quantified using models such as the Underhill-Zucchini moult model. Studies that included estimates of primary moult parameters (duration, mean start date and standard deviation of mean start date) obtained using the Underhill-Zucchini moult model were compiled from the literature to generate a moult analysis database. These parameters provide information about the timing of moult, how it fits into the annual cycle and how synchronised moult is in the population. Relationships between the estimated moult parameters and a series of explanatory variables were explored. Critical gaps in our understanding of primary moult were revealed. The available data are taxonomically biased towards Charadriiformes and Passeriformes and there are substantial gaps in coverage across the phylogenetic tree. There are severe biases in the global distribution of moult studies across continents, with Africa having almost half of all applications of the moult model. Latitude emerged as an important explanatory variable but there are latitudinal gaps in the data, especially from the Tropic of Capricorn extending to 40°N and in the far northern and far southern hemispheres. Citizen scientists, both bird ringers and photographers, have a key role to play in helping to fill these data gaps. An emerging pattern in the available data is that the timing of primary moult is more synchronised in populations in far northern and far southern latitudes than in temperate or tropical zones, but more data in the far south is needed to confirm this. Timing of moult shows great variability in the tropics and sub-tropical regions of the southern hemisphere. In species that undertake an annual sequential moult, larger birds take longer to moult than smaller birds, but the allometric relationship is not tight. Migrants appear to be more synchronised in their moult than residents.

Rapid flight feathers moult and fat stores in the Common Snipe Gallinago gallinago in the early stage of autumn migration

We analysed primary and secondary feather moult and fat reserves in 539 Common Snipes captured in the middle Pripyat River Valley, an important stopover site for waders in Central Europe, between 2002 and 2022. The average daily rate of feather growth was 1.89% in primaries and 2.27% in secondaries, being one of the highest documented in waders. The estimated duration of growth for a single flight feather varied from 11 to 21 days in primaries and from 8 to 11 days in secondaries. Moreover, multiple flight feathers (up to 14) were replaced simultaneously. As a result, the wing moult in Common Snipes was rapid with the mean primary moult duration estimated at 53 days (28 June–20 August) according to the Underhill–Zucchini model, and only 20 days in secondaries (31 July–20 August) based on moult estimates of individual secondaries. Hence, although secondary feathers began to grow when primary moult was already advanced, moulting of both flight feather groups was completed in most birds at almost the same time. Our study shows that Common Snipe in the middle Pripyat River Valley exhibit very rapid wing moult with large wing gaps. Fat reserves and thus body mass of Common Snipes were the lowest when the wing gap was greatest, compensating for their reduced wing area. Late and slow movement towards wintering grounds, allows them to moult rapidly at the early stage of autumn migration, which is likely to occur only in sites with abundant food resources.

Timing and extent of primary moult in breeding Black-legged Kittiwakes Rissa tridactyla at different colonies in the north-east Atlantic

ABSTRACT Timing of moult of the inner primaries by breeding Black-legged Kittiwakes is very similar among colonies ranging from southern England to northern Norway, but can be much later at very high-latitude colonies, such as at Svalbard, where breeding is delayed by the late spring thaw. Breeding adult Black-legged Kittiwakes almost all replaced the first and second (innermost) primary feathers on their wings while breeding.

Second Prebasic Molt of a Black-headed Gull at Anchorage, Alaska

A second-cycle Black-headed Gull (Chroicocephalus ridibundus), a rare to casual visitor in western North America, remained at Anchorage, Alaska, from 16 July through 17 September 2023, providing a unique opportunity to track and document most stages of its second prebasic molt into definitive (adult) basic plumage. I estimated that the Anchorage bird required approximately 3 to 3.5 months (early/mid-June through late September) to complete the molt of its primaries. This is notably longer than the “average” 2.5 months stated for second-cycle Black-headed Gulls in several publications, but consistent with the duration and calendar limits reported by others for the entire species (mid-May/June through September). The start and duration of rectrix molt relative to the stage of primary molt closely matched that described for gulls in general, and specifically for predefinitive Bonaparte’s (C. philadephia) and Western (Larus occidentalis) gulls. When first observed in mid-July, the Anchorage bird had a white tail, as in definitive plumage, a trait infrequent in first-cycle Black-headed Gulls; then it molted the tail again through August. Although Black-headed Gulls occurring in western North America have been presumed to originate from eastern Asia, the timing of the Anchorage bird’s primary molt closely matched that published for second-cycle Black-headed Gulls in Europe, which has reported to be earlier—though molt schedules of Charadriiformes at the population level are highly variable.

When slow explorers are fast: Personality‐related differences in timing of migration in Red Knots (Calidris canutus)

The timing of migration varies significantly among individuals, even within populations sharing breeding sites. Consistent individual behavioural differences, known as personality traits, have been linked to variation in movement behaviour. However, little attention has been given to investigating whether personality traits can explain such variation in the timing of migration. We studied post‐breeding migration of Red Knots Calidris canutus islandica breeding in the High Arctic and wintering in European coastal areas. We captured Red Knots in the Dutch Wadden Sea, a main non‐breeding and moulting site, and assessed their exploration behaviour, a personality trait, before releasing them with colour‐rings. We first investigated whether arrival timing in marine areas was associated with exploration speed. Secondly, we asked whether slow explorers were more likely to fly non‐stop to the Wadden Sea compared with fast explorers, as faster explorers were expected to move more and visit more staging sites. To determine arrival timing in marine non‐breeding areas and non‐stop flights to the Wadden Sea, we analysed isotope signatures in blood samples collected after capture, as these differ between the terrestrial breeding grounds (arthropod diet) and marine non‐breeding and moulting sites (benthic invertebrate diet). Thirdly, we estimated arrival time in the Wadden Sea based on primary moult progress, allowing us to examine the relationship between arrival timing estimated from isotope values and the onset of moult. Our findings revealed that slower exploring Red Knots departed earlier from the breeding site and were more likely to fly non‐stop to the Wadden Sea than were faster exploring individuals. Arrival timing to marine areas as estimated from isotope analyses predicted arrival timing in the Wadden Sea (via moult progress) but this relationship exhibited significant variation, possibly due to individual differences in the use of staging sites en route. By migrating early, slower explorers can better outpace the aerial predators also arriving at the Wadden Sea moulting site and thereby reduce mortality risk due to predation.

From partial to complete: Wing- and tail-feather moult sequence and intensity depend on species, life-cycle stage, and moult completeness in passerines

Passerines moult during various life-cycle stages. Some of these moults involve the retention of a variable quantity of wing and tail feathers. This prompts the question whether these partial moults are just arrested complete moults or follow different processes. To address it, I investigated whether three relevant features remain constant across partial and complete moults: 1) moult sequence (order of activation) within feather tracts (e.g., consecutive outward moult of primaries) and among tracts (e.g., starting with marginal coverts, followed by greater coverts second, tertials, etc.); 2) dynamics of moult intensity (amount of feathers growing along the moult progress); and 3) protection of wing quills by overlapping fully grown feathers. To study the effect of moult completeness on these three features, I classified moults of 435 individuals from 61 species in 3 groups: i) complete and partial, ii) without and iii) with retention of feathers within tracts. To study the effect of life-cycle stage, I used postbreeding, postjuvenile, and prebreeding moults. I calculated phylogenetically corrected means to establish feather-moult sequence within tracts. I applied linear regression to analyse moult sequence among tracts, and polynomial regression to study the dynamics of moult intensity as moult progresses. Sequence and intensity dynamics of partial moults tended resemble those of the complete moult as moult completeness increased. Sequence within and among feather tracts tended to shift as moult intensity within tracts and number of tracts increased. Activation of primaries advanced in relation to the other feather tracts as number of moulted primaries increased. Tertial quills were protected by the innermost greater covert regardless of moult completeness. These findings suggest that moult is a self-organised process that adjusts to the degree of completeness of plumage renewal. However, protection of quills and differences among species and between postjuvenile- and prebreeding-moult sequences also suggest an active control linked to feather function, including protection and signalling.

Using generalised additive models to visualise the annual cycle of primary feather moult patterns

Moult is a vital but energy-demanding event in a bird's phenological cycle. Primary moult is the shedding of old feathers and the regeneration of new flight feathers to maintain efficient functioning of these critical structures. Life history events, such as moult, breeding and migration, are typically separated from one another temporally because of the costs associated with moult Thus, extrapolation of other life-history events is possible from an understanding of the primary moult schedule in species for which data are easier to collect This study aimed to explore the primary moult patterns of nine nectarivorous species (seven Sunbirds and two Sugarbirds) using data extracted from the South African Bird Ringing Unit (SAFRING) database. Logistic generalised additive models (GAMs) were fitted to active/inactive moult observations to explore annual primary moult patterns. Breeding patterns based on the presence of brood patches in females were also analysed using GAMs for species with sufficient data. The GAMs illustrated the annual patterns well, with the probability of moult having clear peaks and troughs, allowing intuitive understanding of annual life-history patterns, such as the probability of breeding. Generally, GAMs hold potential to be used as a supplementary tool to study the timing of primary moult GAMs appear especially useful for species with no well-defined annual moult period.

Sea ice concentration decline in an important Adélie penguin molt area.

Unlike in many polar regions, the spatial extent and duration of the sea ice season have increased in the Ross Sea sector of the Southern Ocean during the satellite era. Simultaneously, populations of Adélie penguins, a sea ice obligate, have been stable or increasing in the region. Relationships between Adélie penguin population growth and sea ice concentration (SIC) are complex, with sea ice driving different, sometimes contrasting, demographic patterns. Adélie penguins undergo a complete molt annually, replacing all their feathers while fasting shortly after the breeding season. Unlike most penguin species, a majority of Adélies are thought to molt on sea ice, away from the breeding colonies, which makes this period particularly difficult to study. Here, we evaluate the hypothesis that persistent areas of high SIC provide an important molting habitat for Adélie penguins. We analyzed data from geolocating dive recorders deployed year-round on 195 adult penguins at two colonies in the Ross Sea from 2017 to 2019. We identified molt by detecting extended gaps in postbreeding diving activity and used associated locations to define two key molting areas. Remotely sensed data indicated that SIC during molt was anomalously low during the study and has declined in the primary molt area since 1980. Further, annual return rates of penguins to breeding colonies were positively correlated with SIC in the molt areas over 20 y. Together these results suggest that sea ice conditions during Adélie penguin molt may represent a previously underappreciated annual bottleneck for adult survival.

Primary moult in the annual cycle of adult African Oystercatchers Haematopus moquini

African Oystercatchers Haematopus moquini are sedentary, intertidal foraging birds, range-restricted to Namibia and South Africa. Their annual cycle has two components: breeding and moult. The Underhill–Zucchini model estimated a primary moult duration of 154 days, and mean start and completion dates of 1 April and 2 September. Most eggs (90%) are laid between mid-November and mid-February; thus, on average, oystercatchers have 2–3 months between moult completion and egg laying. The completion date confirms an overlap between moult and the period of post-fledging dependence of juveniles on their parents. The 154-day primary moult duration spreads the energetic demands of feather production to cope with this overlap and with the fact that primary moult extends through winter when storms make intertidal areas inaccessible for days. There are approximately 50 quantitative studies of primary moult for migrant waders (suborder Charadrii); this is the third for a resident species. The other two species of resident waders studied are in south-eastern Australia: the Hooded Dotterel Thinornis cucullatus and Sooty Oystercatcher Haematopus fuliginosus. The Hooded Dotterel has one of the longest estimated moult durations (203 days) and its moult and breeding overlap. The annual schedules for the Sooty Oystercatcher and African Oystercatcher are similar, as is the estimated moult duration: 149 and 154 days, respectively. Preliminary studies of the primary moult of the Eurasian Oystercatcher H. ostralegus indicate the duration to be 100 days, which is two-thirds that of the two sedentary oystercatchers. This is likely a consequence of Eurasian Oystercatchers having migration and harsh winter conditions as additional components of their annual cycle, and thus a limited period for primary moult. This study was based largely on photographic records of African Oystercatchers in flight, with contributions by citizen scientists. We make two key conclusions: the need for more studies of the moult strategies of the Haematopidae and resident waders; and digital photography of free-ranging birds can be used to study primary moult.

The primary moult of adult Red-headed Finches Amadina erythrocephala

Little is known about the biology of the Red-headed Finch Amadina erythrocephala. This study investigates its primary moult, the first aspect of its life history to be looked at in depth. Additionally, we used a modified Cook’s distance algorithm to explore the influence of individual years on the moult parameter estimates. Lastly, the primary moult of Red-headed Finches is compared to that of other passerines resident in the Highveld region of South Africa. Red-headed Finches breed through the Highveld winter, and start moult on 22 August, on average. Moult extends through spring and most of the summer, reaching completion on 22 February, on average. However, the moult started unusually early in the years 1988, 2005, 2008 and 2019, whereas in 2017 moult started unusually late. The duration of moult was estimated to be 185 days. The majority of other passerines resident in the central Highveld breed through summer, undergo primary moult in autumn, and complete moult before the regular frosts start to occur during the Highveld winter. The moult duration of the Red-headed Finch is substantially longer than that of other resident passerines.

Pattern and intensity of moult in White-capped/Shy Albatrosses Thalassarche steadi/cauta

Albatrosses typically only replace a subset of flight feathers each annual moult cycle. We scored the moult of 662 White-capped/Shy Albatrosses Thalassarche steadi/cauta, using birds that had been killed by longline fisheries off South Africa. Most adults and immatures alternated replacing the outer primaries (phase 1) and inner primaries (phase 2), but there was considerable variation among adults, with 0–10 primaries replaced each moult cycle. Birds in phase 1 moult replaced fewer primaries than birds in phase 2, but the proportion by mass was similar. Adults replaced more primaries than immatures as there was a disproportionate number of adults in phase 2 moult (82%), which might indicate that most adults off South Africa are in their first year after a successful breeding attempt, but a similar bias occurred among beached birds from New Zealand and Australia. The outer primaries were usually replaced outwards from P8, but the replacement sequence was more variable among the inner primaries. Secondary moult overlapped with primary moult, but not all secondaries were moulted each year. Adults had more active moult centres and replaced more secondaries at once than immatures. Tail moult started after primary moult, with 1–11 feathers growing from 1–6 active centres. Some birds replaced their rectrices in pairs, with often 2 or 3 pairs moulting at once, but others replaced alternate feathers or almost all rectrices at once. A few birds exhibited intense wing moult, but the norm was to replace only 1 or 2 primaries and 1–6 secondaries at once. Age- and sex-related differences in moult intensity presumably result from different time constraints. Adult White-capped/Shy Albatrosses take a year off between successful breeding attempts, allowing time for a more protracted moult. However, we cannot explain the intense moult in some immatures. Our findings support recent studies showing that albatross flight feather moult patterns are more varied than previously reported.

Primary moult of adult Laughing Doves Spilopelia senegalensis: introduction of the relative duration index and a comparison of moult indices

A new moult index, the relative duration index (RDI), was developed to quantify progress through moult using the Underhill–Zucchini model of primary avian moult, with the Laughing Dove Spilopelia senegalensis as a case study. The RDI weights individual primaries in accordance with their estimated relative moult durations in the same way that proportion feather mass grown (PFMG) weights primaries in accordance with their relative masses. A sample of 1 893 moult records of adult Laughing Doves was used to study their primary moult. Moult parameters were estimated for three moult indices: the traditional moult score (TMS), PFMG and the RDI, and for each of the 10 primaries. Comparing the moult parameter estimates of each of the three indices to their respective cumulative growth curves obtained from the progression of moult of individual primaries, PFMG had the closest fit. However, any of the three moult indices would give satisfactory results for the Laughing Dove. Laughing Doves had an estimated primary moult duration of 215 days, with mean start and end dates of 3 November and 6 June, respectively. The standard deviation parameter was 66 days, which implies a lack of synchronisation of moult in adult doves. It was hypothesised that the RDI might be an appropriate moult index for species that moult one feather at a time and for which the individual moult durations are not proportional to their mass, as seen in Laughing Doves. However, this was not the case as PFMG provided the best fit. The RDI might also prove to be an appropriate moult index for species that moult multiple feathers simultaneously during part of the moult period and these feathers grow more slowly than when fewer feathers are moulting simultaneously.

Moult performance varies in relation to colour patterns in crossbills

Carotenoid-based ornamentation can vary greatly among individual birds of the same population. This variability might consider the timing and duration of moult. Crossbills (genus Loxia) show large variation in their colour patterns, ranging from dull yellow to red. Thus, they provide an excellent avian model for testing whether the timing and duration of their moult are associated with their dominant colour patterns. Using a dataset of more than 1900 crossbills captured in the Pyrenees (Spain), we observed that individuals with red feathers started their primary moult early, while those with yellow feathers started moult on an average of 18 days later. We also found that yellow crossbills were more likely to suspend moults (i.e. temporally interrupt moult), postponing a part of the moult until September. These differences in the moulting process may be related to plumage colour, which reflect individual condition.

Using banding data to estimate primary feather molt and hatch chronology of white‐winged doves in Texas

AbstractBanding data are commonly used to estimate vital rates for migratory game bird management. We used white‐winged dove (Zenaida asiatica) banding data to estimate molt and hatch chronology in Texas. We used Texas Parks and Wildlife Department's long‐term, state‐wide banding data (71,675 banded individuals) from 1 June to 15 August 2007–2016 to investigate primary feather molt and hatching in white‐winged doves in Texas. We estimated primary feather molt and used individual recapture data to determine reliability of models predicting primary feather molt rates. For hatching, we used primary feather molt scores of captured hatch‐year doves to backdate to an estimated hatch date. Our modeling predicted mean after‐hatch‐year primary feather molt rate of 13.21 ± 0.93 days. We predicted 95% of adult white‐winged doves began molting between 7 April to 8 July and completed molt between 17 August to 17 November. Across all years, white‐winged doves hatched as early as 6 January and as late as 27 July, with 95% of all hatching occurring between 22 March and 18 June and peaking on 4 May. Primary feather molt initiation peaked 16 days after the peak of hatching, suggesting that white‐winged doves delay the onset of primary molt until reproductive activity slows. Secondary data collected during banding operations on migratory game birds may be used to understand additional life processes without the requirement to initiate additional survey efforts.

Biometrics and moult of Grey-backed Sparrow-lark Eremopterix verticalis in the Karoo, September 2023

The biometrics of Grey-backed Sparrow-larks Eremopterix verticalis in the handbooks are based on small samples. We summarize wing-length, tail-length and mass of a sample of 130 Grey-backed Sparrow-larks mistnetted during ringing on the farm New Holme, Hanover District, Northern Cape, South Africa, in September 2023. The habitat was Grassy Karoo with, on average, March being the month of peak annual rainfall. The sample consisted of 58 females and 72 males. We examined birds for primary moult; 5% had moulted either one or two primaries, indicating that moult was commencing in the population as a whole. This observation, set in the context of an observation made in Namibia, suggests that this species undertakes a pre-breeding moult in spring and summer, prior to breeding in late summer and autumn.

Plumage dimorphism and moult patterns of weaverbirds – a review

The colourful plumages of weaverbirds (family Ploceidae, with 118 species recognised) provide a well-known and in part iconic showcase of sexual selection, with widespread sexual dimorphism (93 species) as well as seasonal dimorphism (62 species). The tremendous diversity in both the type and degree of plumage dimorphism is poorly known, which calls for a better and extended understanding of the highly variable moult patterns in this group. Here, we lay a foundation for this by reviewing and discussing what is known of the timings and patterns of moult in the Ploceidae. The extent of post-juvenile moult is poorly known in weavers: it is known to be complete in 10 species, partial in two species, and either complete or partial in six species. However, it is likely that there is more variation than previously thought. Weavers typically have a complete moult after breeding, and before breeding a partial prenuptial moult, restoring the male (and, in some species, female) breeding plumage; this usually involves only body feathering, one exception being the Seychelles Fody Foudia sechellarum, which has a complete prenuptial moult. We compiled primary moult data based on the proportion of feather mass grown, for 11 weaver species, three of which are newly reported on. Timing of primary moult is discussed in relation to sex, location and annual timing. The duration of primary moult varies widely between and within species, from 46 days (1.5 months) in a widowbird to 215 days (7 months) in an arid species. Weavers in arid regions have longer moult durations than weavers in other regions. Examples of the overlap of primary moult and breeding are presented.

The relationship of moult timing, duration and sequence to the aerial lifestyle of the Little Swift (Apus affinis)

Feather moult is an important and energy‐demanding avian life‐history trait that is necessary to maintain the function of the plumage by replacing abraded feathers with new ones. However, during moult, aerodynamic capacity is temporarily impaired by the reduction and shape change of the flight surface. Therefore, among aerial birds, adaptations may evolve to reduce this impairment. Here, we studied the timing, duration and sequence of the annual complete moult of adult Little Swifts Apus affinis in northern Israel by examining the plumage of 55 individuals monthly in their colony. Primary moult occurs over 191.7 ± 24.0 days, representing 52.5 ± 6.6% of the annual cycle, and overlaps with the breeding season, a rare occurrence among small Palaearctic bird species. In addition, primary moult duration was significantly affected by its timing such that individuals that started to moult later had a shorter moult duration than those that started to moult earlier. This relationship is most likely to be a result of the time available for moult, which is limited by breeding and also by the cold season and the expected decline in food availability. Moult overlap with breeding or cold seasons may involve energetic and functional costs. Our results indicate two strategies that appear sequentially: (1) a slow and long moult that largely overlaps with breeding, and (2) a shorter moult that overlaps less with breeding (both strategies avoid moulting during the cold season). Most Little Swifts (85.7%) moulted their rectrices in a centripetal sequence starting with the outermost pair of rectrices inward. This moult was later in relation to wing feather renewal and had almost no overlap with primary moult. Finally, we tested a method for non‐invasively monitoring moult of small species by collecting and sampling shed feathers without the need to capture the birds themselves, as has been successfully used for larger bird species. This work demonstrates the importance of aerodynamic considerations in the evolution of life‐history traits. Moult‐related aerodynamic costs may be an important evolutionary factor shaping moult strategy, including timing, duration and sequence, mainly among species, like swifts, that are highly dependent on their aerodynamic ability.

Wing-moult and seasonal occurrence of sunbirds (Nectariniidae) captured at four sites in the Eastern Cape, South Africa

Ringing activity in the Eastern Cape province, South Africa, spanned 30 years, with three sites active concurrently over a 14-year period, and three sites over a 6-year period. In total, in the course of 714 ringing days, in all months, we handled 1 207 Amethyst Sunbirds, 708 Southern Double-collared Sunbirds, 707 Greater Double-collared Sunbirds, 381 Grey Sunbirds, 121 Malachite Sunbirds and 140 Collared Sunbirds. Although individuals of all six species were recaptured more than five years after ringing, and despite evidence of the seasonal passage of sunbirds at two sites near the Indian Ocean coast, no birds were recaptured away from the original ringing site. The extent of seasonal movement by these species remains unclear. All species have a complete post-breeding moult, with a period of eclipse plumage in males of the Malachite Sunbird and Southern Double-collared Sunbird. Primary wing-moult in adult Greater and Southern Double-collared Sunbirds extends over three months beginning in December. Inland populations of adult Amethyst Sunbirds and Malachite Sunbirds start primary moult in April. Moult apparently proceeds more rapidly in Malachite Sunbirds (70 days to completion) than in Amethyst Sunbirds (125 days), but the Malachite Sunbird is the only species that frequently displays interrupted/suspended moult in this region. Adult Amethyst Sunbirds in the coastal population evidently start moulting earlier, in late February, with a similar duration to completion as inland birds. Juvenile Amethyst Sunbirds start to moult some two months later than adults in each region and they may complete moult more quickly. Collared Sunbirds appear to have the most protracted moult, extending over five months from November, but estimates of moult parameters for this species and the Grey Sunbird (three months from December) are based on small samples.