Feathers made of keratin are structures that are subject to wear and tear. As proper performance lasts less than a year for most feather types, birds replace them regularly, often once or twice a year, a body maintenance process called moult (Payne 1972, Jenni & Winkler 2020). The growth of feathers is a resource-demanding process (Murphy 1996, review in Jenni & Winkler 2020) and as birds with actively growing feathers may suffer functional handicaps with respect to insulation, ornamentation or flight (review in Jenni & Winkler 2020), the timing of moult of various feather tracts should be tightly integrated with other annual cycle challenges such as breeding and migration (e.g. Kjellén 1994, Buehler & Piersma 2008, Dietz et al. 2013). For example, migrants tend to separate moult, even of their body feathers, from long-distance migratory flights (Piersma & Jukema 1990, Zwarts et al. 1990, Jenni & Winkler 2020). In large species this poses problems because chick rearing time is long, and their large feathers require a disproportionately long time to grow (Hedenström 2006, Rohwer et al. 2009). As the length of a year is fixed, moult may be suspended before migration and resumed afterwards (Ginn & Melville 1983, review in Jenni & Winkler 2020). In such suspended moult, some feathers are completely replaced while others are left for later renewal, resulting in fully-grown wings consisting of two (and sometimes more) feather generations.

As discussed by Rohwer et al. (2009) and Jenni & Winkler (2020), larger birds have a hard time to fit even a single (and possibly suspended) descendent moult into an annual cycle. Such species may solve the problem by simultaneous or synchronous primary moult (in grebes Podicipedidae, flamingos Phoenicopteridae, waterfowl Anseriformes and cranes Gruidae; e.g. Piersma 1988, Jehl 1990), or breaking the primary moult up into stages, where moult is suspended at some point and resumed to complete it either within a year, or in multiple years. Rohwer et al. (2009: Figure 1C) showed that only three of c. 23 reviewed species heavier than 1500 g managed to complete primary moult without losing the capacity to fly.

More than 50 years ago, in their grand review of moult, Stresemann & Stresemann (1966) stated that Eurasian Spoonbills Platalea l. leucorodia (henceforth called ‘spoonbills’), birds weighing up to 2 kg, would replace their outer flight feathers, the primaries, in a descendent way, i.e. that the primaries would be replaced from the innermost out, leaving a gap in the wing where new primaries are growing. Nevertheless, Stresemann & Stresemann (1966) added a question mark to the statement and it remains unclear on which data their inference is based. Bauer & Glutz von Blotzheim (1966) summarized the moult of spoonbills as follows: “Altvögel scheinen sowohl in Spätsommer und Früherbst als auch von Januar bis März im Winterquartier das Gesamte Gefieder zu wechseln, also im Laufe eines Jahr zwei Vollmausern durchzumachen. Eingehendere Untersuchung der Mauser ist dringend erwünscht”. (“Old birds seem to change their entire plumage in late summer and early autumn as well as from January to March in the winter quarters, i.e. they go through two complete moults in the course of a year. A more detailed study of the moult is urgently desired.”) Thirty years later, Cramp & Simmons (1997) added some detail: “Wing moult starts with inner primaries in August–September; probably stops [i.e. is suspended] during migration, and is finished in winter quarters”. Primary moult patterns of other species of spoonbill such as the Yellow-billed Spoonbill Platalea flavipes and the Roseate Spoonbill Platalea ajaja, remain largely undescribed as well (Dumas 2020, Matheu et al. 2020). Furthermore, age-related differences in moult patterns of spoonbills are not described at all.

We examined the primary moult of spoonbills in the eastern Dutch Wadden Sea from July to October 2021, using photos collected on Schiermonnikoog where spoonbills breed and from Lutjewad, on the mainland coast of the province of Groningen, a site where large numbers of Schiermonnikoog-breeders are joined by other spoonbills after the breeding season. Lutjewad is used as a staging site prior to the southward migration towards wintering grounds ranging from France to Senegal (Piersma et al. 2022). We aim to verify and expand the limited knowledge of the primary moult of spoonbills in quantitative ways using digital photography. Of particular interest are age differences in the onset and speed of primary moult, as well as the suspension of primary moult and its overlap with fuelling prior to southward migration. Assuming that the onset of moult is constrained by breeding activities, we expect non-breeding immatures to initiate moult several weeks before adults (Boere 1976). However, due to their greater experience and knowledge of resources in the Wadden Sea, adults may be able to moult faster, and achieve similar moult scores as immatures prior to migration.

METHODS

The study area in the eastern part of the Dutch Wadden Sea consisted of the system of intertidal flats and gullies bordered by the barrier island Schiermonnikoog in the north and the dike of the Dutch mainland, including the embanked Lauwersmeer in the south (Figure 1). Throughout summer, this entire area is used by spoonbills breeding on Schiermonnikoog (e.g. Lok et al. 2023). After the fledging of young, spoonbills gradually leave the island for the mainland coast and the Lutjewad area in the province of Groningen appears to be of particular significance (Figure 1; Piersma et al. 2022). Accounting for these individual movements, to best ‘capture’ the entire primary moult cycle, observations were collected both at Schiermonnikoog and at the Groningen coast (Figure 1).

Figure 1.

Study area of Schiermonnikoog and Lutjewad in the eastern Dutch Wadden Sea, along with examples of shifts of individually colour-ringed Eurasian Spoonbills from Schiermonnikoog to Lutjewad during July–September 2021 (top panel), and the distribution of observation dates of individually colour-ringed spoonbills at Schiermonnnikoog and Lutjewad depending on whether they were seen at one or both study sites. The boxes enclose 50% of the data points, the vertical lines show the median date and the horizontal lines show the range of observation dates. Note that at Lutjewad observations only started in mid-July.

At Schiermonnikoog (53°28′N, 06°10′E) photographic sampling of spoonbills passing by in flight, at close or medium range, was performed every few days from 27 June to 1 October 2021. Most photos were taken as birds were flying in and out of a small lake, the Westerplas, and across the marina area in the southwest of Schiermonnikoog. On the Groningen coast off Hornhuizen (53°24′N, 06°20′E), just south of the low tide feeding grounds of Lutjewad (Piersma et al. 2022), 11 full field days were spent from 17 July to 27 September. Here, next to a high tide roost for waterbirds, we set up a hide four hours prior to high tide to avoid disturbance to the roosting birds. Roughly four hours after high tide, after all spoonbills and other waterbirds had left on their own accord, the hide was dismantled and the study area left. At Schiermonnikoog a Nikon D500 body with a 300-mm f/4.0 lens was used, while at Lutjewad we used a Canon 7D Mark ii body with a 400-mm f/5.6 L lens.

Over 17,000 images were processed and examined using Adobe Photoshop Lightroom Classic CC. Minor contrast and lighting adjustments were occasionally applied to enhance visibility of primary feathers or growing primary pins. A total of 1471 usable images, each one representing a single data point, were obtained, 535 of which were collected on Schiermonnikoog and 936 on Lutjewad. 513 of these images were of juveniles hatched earlier that summer, recognizable by their light and fleshy coloured bills of relatively small size, and none of them showed any traces of primary moult. After excluding these images, a total of 958 images of non-juvenile spoonbills were used in the analysis.

Figure 2.

Typical field images used to score primary moult of Eurasian Spoonbills. On the left are three example images of anonymous individuals (photos by A.P. de Boer), on the right three images showing progression of moult in Sinagote (photos top to bottom by M. Muller, A.P. de Boer and S.K. de Vries), a female breeding on Vlieland and moving for moult to Lutjewad (see Piersma et al. 2022). PS = Primary Score.

Birds in flight were aged as ‘immature’ on the basis of black-coloured tips of especially the outermost primary flight feathers (Figure 3; for a review of this and other ageing criteria, see A. Hellquist at  http://surfbirds.com/mb/Features/spoonbill/ageingspoonbill-0402.html). Birds in their 2^nd and 3^rd calendar-year which are unlikely to have been breeding are called ‘immatures’. Birds with slightly mottled or completely white outer primaries were considered ‘adults’ (Figure 3).

High-quality photos, i.e. photos that were sharp and clearly showed all 10 primaries of at least one wing (Figure 2), were selected for scoring, for which we used the system widely used for birds in the hand (Ginn & Melville 1983). Here, each of the 10 primary feathers received a score ranging from 0 to 5. Old feathers (showing wear or discoloration, score 0), a missing/shed feather (score 1), up to a third grown (score 2), up to two-thirds grown (score 3), up to completely grown (score 4) or completely grown (score 5). The resulting primary score was the sum of all primary feather scores, ranging from 0 to 50. As the scores 1–4 indicate growing feathers, moult suspension occurred when we observed primaries with scores of 0 and 5 in the same wing.

Jenni & Winkler (2020), on the basis of an in-depth review, conclude that the shedding intervals of primaries are such that a constant rate of feather material production is achieved. This results in the outer primaries of spoonbills, which are longer than the inner primaries, taking more time to fully regrow. To linearize the process of moult progression, we determined the proportions of feather mass grown (PFMG), ranging from 0 to 1 (Underhill & Joubert 1995). A set of slightly worn primaries was obtained from a deceased spoonbill and the dry mass of each primary weighed up to the nearest mg; the following averages of three measurements were used: P1 = 0.441 g, P2 = 0.473 g, P3 = 0.515 g, P4 = 0.591 g, P5 = 0.665 g, P6 = 0.793 g, P7 = 0.928 g, P8 = 0.925 g, P9 = 0.897 g and P10 = 0.757 g. The proportion of the feather mass grown in relation to the total mass of all primary feathers was calculated using the function ‘ms2pfmg’ from the R-package ‘moult’ (Erni et al. 2013).

Figure 3.

Black wing tips of Eurasian Spoonbills. The extent of black on the tips of the outer primaries decreases with age and are approximated to calendar year (CY) as follows: (A) Juvenile, (B) 2CY, (C) 3CY, (D) 4CY, (E) Adult (photos by A.P. de Boer).

Figure 4.

The scoring of the abdomens of Eurasian Spoonbills in flight, from 1 (lean) to 5 (fat), inspired by methods routinely used in studies of geese (Owen 1981) and shorebirds (Wiersma & Piersma 1995; photos by A.P. de Boer).

To assess the body condition of spoonbills we used a 5-score abdominal profile index, as developed by Owen (1981) for the study of geese, and also applied to much smaller shorebirds (Wiersma & Piersma 1995). The abdominal profile of standing and walking spoonbills appeared impossible to score sensibly, possibly due to the often fluffed belly feathers. However, in flight the belly feathers seem to be held rather tightly to the body, and we therefore scored the abdominal profile (ranging from 1 = lean to 5 = fat) for birds in flight (Figure 4). Of the 958 images, 880 were suitable for abdominal profile scoring.

The number of repeat observations in our dataset is unknown, but not negligible. Although a substantial portion of birds using the area were colour-ringed, the colour ring combinations were not always completely legible on images of birds in flight. 36 colour ring combinations of adults and immatures were read on the photographs, of which six were observed more than once a single day, and 11 of which were observed more than once throughout the observation period. A more easily recognizable individual in our images was ‘Sinagote’, a female with a rich history of observation (Piersma et al. 2022) carrying a black tracker on her back (Figure 2). From her tracking data we know that Sinagote was present at Lutjewad from 21 July to 14 September. During this period we had eight field days at Lutjewad and photographed Sinagote on four of these days. This suggests a substantial portion of repeat observations in our dataset.

All analyses were conducted in statistical software R v. 4.0.2 (R Core Team 2020). All graphs were produced using the ‘ggplot2’ package (Wickham 2016). We first explored the general pattern of moult progression visually, by comparing running averages of moult scores with a loess-smoothing procedure in ggplot2. This revealed a sigmoidal trend, indicating population turnover at the start and end of our study period. Moreover, our data indicate that our study period did not cover the beginning of the moult period, especially for immatures, and the completion of moult (in both age groups). Therefore, to estimate the start dates and rate of moult of different age cohorts we used type 3 Underhill-Zucchini models (Underhill & Zucchini 1988), which are implemented by the function ‘moult’ of the R-package ‘moult’ (Erni et al. 2013). The type 3 model uses only actively moulting individuals (i.e. PFMG score > 0 and < 1).

Figure 5.

Seasonal progress of primary moult of immature (left) and adult (right) Eurasian Spoonbills expressed as: (A,B) the number of the last shed primary, (C,D) the primary moult score and (E,F) the Proportion Feather Mass Grown (PFMG). The data points for the individual called ‘Sinagote’ are marked in green. The blue curves are running averages with a 95% confidence interval to describe seasonal patterns. The extrapolations from the UZ models are shown as red lines (E&F).

RESULTS

The primary scores of spoonbills from the eastern Dutch Wadden Sea increased over time, with some immatures achieving scores of 20–30 (i.e. 4–6 newly grown primaries), and some adult scores up to 10–12 (i.e. 2 newly grown primaries) in late June (Figure 5). Few individuals completed primary moult in our study area.

As anticipated on the basis of individuals moving from Schiermonnikoog to Lutjewad in the course of the study period (Figure 1), the overlapping data distributions from the two sites are consistent with the idea that we were sampling the same population at both locations. The progress in primary moult scores was not linear over time, regardless of whether it was measured in terms of the last shed primary, primary score, or PFMG (Figure 5). Running averages computed over our full dataset (including birds that did not yet start moult and those that completed moult), showed clear sigmoidal trends. By contrast, the steady, linear progress of primary moult of the repeatedly photographed individual Sinagote nicely paralleled the steepest slope of the running average during the middle of our study period (Figure 5B, D, F). According to type-3 Underhill-Zucchini models (Table 1), immatures would need 131 days to grow all primaries (starting on 11 June) whereas adults would require 170 days (starting on 20 July), without any suspension. We cannot estimate the real end date of moult, because this would depend on how long it is suspended during and after migration.

Figure 6.

Seasonal change, at weekly intervals from 27 June, in the incidence of the suspended moult in immatures (left) and adult (right) Eurasian Spoonbills (top row) and the distributions of their abdominal scores (bottom row).

Table 1.

Moult parameters for immature and adult Eurasian Spoonbills according to the Underhill-Zucchini method.

The proportion of scored spoonbills with suspended primary moult in Lutjewad and on Schiermonnikoog was rather low, rising sharply only from September onwards to 35% of immatures and 40% of adults by late September (Figure 6A). Only three immatures out of 168 samples were scored as having completed primary moult and only one out of 790 samples for adults. Lean birds were only found until September, but fat birds were seen throughout the study period (Figure 6B), with immature birds slightly lagging behind adults (Figure 6C). As such, spoonbills' abdominal profiles increased well before they suspended their moult.

DISCUSSION

The observations reported here confirm the inferences by both Stresemann & Stresemann (1966) and Bauer & Glutz von Blotzheim (1966) that Eurasian Spoonbills show a descendent primary moult. Very few birds, however, completed primary moult in the Wadden Sea and only 35–40% of birds observed in September suspended primary moult before migrating. We suspect that most Spoonbills leave their moulting area as soon as moult is suspended and are therefore underrepresented in our sample. In addition, based on post-breeding movements of GPS-tracked Spoonbills (unpublished data), it seems likely that at least some of the birds that initiated moult at our study site moved to other nearby staging sites before suspending moult and initiating migration. Moult suspension is consistent with the suggestion in Cramp & Simmons (1997) that primary moult stops (i.e. is suspended) during southward migration, so that spoonbills would have to complete the moult in their winter quarters. Thus, we have been able to empirically underpin the sparse existing descriptions of spoonbill primary moult.

Furthermore, adults started primary moult about five weeks later than immatures. This was expected based on the idea that breeding and moult will be separated in time in most species (Boere 1976). This is certainly the case for Sinagote, a female who, after having tended chicks, was photographed on Vlieland on 13 July without any signs of moult. She was later photographed at Lutjewad, c. 90 km to the ENE, on 1 August with a primary score of 13 and P4 being the last shed primary and on 25 August with a primary score of 25 and P6 being the last shed primary.

Our results also confirm that moult is separated in time from migration in spoonbills. Indeed, the incidence of suspended moult sharply increased among birds of both age groups before they all left the summer range by late September. We also demonstrated that fuelling for southward migration completely overlapped with the moult of the inner half of the primaries in both age groups. Spoonbills thus avoided migrating with large gaps in the wing and seem to synchronize the suspension of post-breeding moult and fuelling between age groups.

Did we discover anything new? We may have, by showing that immatures not only start primary moult earlier, but also moult faster than adults according to the Underhill-Zucchini models. This allowed non-breeding immatures to replace 2–3 additional primaries (Figure 5) compared to adults, before the onset of migration. If continued without suspension, primary moult would be 39 days shorter in immatures than in adult birds; 131 days (a little over four months) as opposed to 170 days (five and a half months). On the basis of the scaling relationship presented by Hedenström (2006): T_moult (d) = 34.4 × body mass^0.19 (g), the primary moult duration of a 1850-g spoonbill would be predicted at 144 days, which is considerably shorter than our estimate for adults and slightly longer than we estimated for immatures. This discrepancy could be due in part to imperfections in our study design, such as population turn-over during our study period. However, our estimates correspond strongly with a recent review by Scott & Underhill (2024, see Figures 9 and 10), as well as with the moult speed of Sinagote, which paralleled the steepest slope of the running average for adults in the middle of our study period, suggesting the discrepancy is due to other reasons.

The reviews of Dietz et al. (2015) and Jackson & Underhill (2022) suggest that if spoonbills moult like shorebirds, moult rates at more southerly latitudes might well be lower than in the Wadden Sea. Documenting differences in moult speed between sites would provide an interesting case of it being a function of location rather than of taxon only. The pattern of the suspended primary moult suggested in our study differs from the complete descendent moult shown by most shorebirds in the Wadden Sea including Red Knots Calidris canutus and Bar-tailed Godwits Limosa lapponica (Boere 1976, Dietz et al. 2013). Instead, the suspended primary moult, such as that of spoonbills, is characteristic of at least two inland shorebirds, i.e. Black-tailed Godwits Limosa l. limosa (van Dijk 1980) and Ruff Calidris pugnax (Koopman 1986).

As adults have less time to replace the primaries (between breeding and migration) than young non-breeders, it may come as a surprise that the moult speed of experienced adults is slower than that of immatures. Then again, slower feather growth tends to result in higher quality feathers (Dawson et al. 2000, Hera et al. 2009); according to Scott & Underhill (2024), a moult duration of 170 days would lead to ‘good quality’ feathers for a bird the size of a Spoonbill, while 131 days would be placed in the ‘poor quality’ category. This means that the faster moult of immatures likely results in feathers of lower quality, which may shed a new light on the fact that immature spoonbills grow their outermost primaries with a black tip and all primaries with a black rachis, a physical trait common to all spoonbill species (del Hoyo 1992). Black keratin is stronger than white keratin (Jenni & Winkler 2020). Thus, black-tipped primaries of juveniles and immatures may represent a compensatory investment in feather strength. Such a ‘protective cap’ may also help young spoonbills to cover longer periods with a single set of primaries compared to adults. After all, juvenile feathers grown in May–July will need to last at least until June–Sept of the next year, i.e. 14 months. Some outermost primaries need to last even longer until after the second southbound migration in Sept–Oct of their second calendar year, i.e. 15–16 months or more. The ‘operational time’ of flight feathers will ultimately decrease to roughly 12 months in adult spoonbills.

In spoonbills photographed in flight during southward migration, e.g. during their crossing from Spain to Morocco (Piersma et al. 2020), we were unable to find birds with moult gaps. This means that we now need to know where and when the spoonbills that started primary moult in the Wadden Sea complete it. If they would resume moult upon arrival on their individual wintering grounds (Lok et al. 2011), we expect to find spoonbills actively moulting primaries in October–November from Atlantic France in the north to Senegal in the south. Alternatively, if they resemble the closely related ibises (Threskiornithinae), spoonbills may resume moult only in the latter stages of the non-breeding period before spring migration (Davis & Kricher 2020). Either way, we expect Eurasian Spoonbills to complete their descendent moult within a single non-breeding period.

All considered, we think our most significant finding is that immature and adult spoonbills synchronize the suspension of moult ahead of migration, regardless of marked age differences in the onset and rate of primary moult, with more new primaries in place in immatures than in adults. For immatures and adults with prior migratory experience this synchronization is probably not essential for learning viable routes and non-breeding destinations (Mueller et al. 2013, Aikens et al. 2022, Lok et al. in prep). However, there is ample evidence to suggest that synchronizing (pre-)migratory activities may also allow experienced spoonbills to achieve safer and more efficient long-distance flights, e.g. through collective navigation (Guttal & Couzin 2010, Berdahl et al. 2018) and formation flight (Weimerskirch et al. 2001, Portugal et al. 2014). The next vanguard question is how they achieve this synchronization. The recent discovery of frequent vocal communication at post-breeding and stop-over sites certainly suggests that spoonbills actively deliberate migratory departures (Lagarde & Piersma 2021).