The path to adult dress: primary moult in second-year Wood Sandpipers Tringa glareola in southern Africa MAGDALENA REMISIEWICZ1,2, ANTHONY J. TREE3, LES G. UNDERHILL1,4 & P. BARRY TAYLOR5 1 Animal Demography Unit, Department of Zoology, University of Cape Town, Rondebosch 7701, South Africa. biomr@ug.edu.pl 2 Avian Ecophysiology Unit, Dept. of Vertebrate Ecology and Zoology, University of Gdańsk, al. Legionów 9, 80-441 Gdańsk, Poland 3 Zoology Department, Nelson Mandela Metropole University, PO Box 1600, Port Elizabeth 6000, South Africa 4 Marine Research Institute, University of Cape Town, Rondebosch 7701, South Africa 5 School of Biological and Conservation Sciences, University of KwaZulu-Natal, Private Bag X01, Scottsville 3209, Pietermaritzburg, South Africa Remisiewicz, M., Tree, A.J., Underhill, L.G. & Taylor, P.B. 2010. The path to adult dress: primary moult in second-year Wood Sandpipers Tringa glareola in southern Africa. Wader Study Group Bull. 117(1): 35–40. Keywords: Wood Sandpiper, Tringa glareola, primary moult, Underhill–Zucchini model, southern Africa We know little about the primary moult of waders in their second year of life, especially migrants. Remisiewicz et al. (2009, 2010) have provided details on the primary moult of immature and adult Wood Sandpipers Tringa glareola in southern Africa, but there is no information on the primary moult of second-year birds. Most Wood Sandpipers leave southern Africa for their northern breeding grounds when they are 10–11 months old, so migration separates the subsequent cycles in their primary moult. We chose this species to determine if the pattern of the irst complete primary moult of waders during their second year of life differs from that of adults. We analysed the primary moult scores of 97 sub-adult (13- to 20-months-old) Wood Sandpipers obtained in southern Africa by using the Underhill–Zucchini moult model to estimate the timing and duration of moult for all 10 primaries combined and for P1 and P2 individually. Sub-adult Wood Sandpipers were observed in southern Africa between June and December, when, by about 19 months of age, they become indistinguishable from adults. Half of the sub-adults showed two generations of their fully grown primaries after a previous partial moult. All 54 sub-adults in active moult started at P1 and progressed outwards to P10. The starting date of moult for all sub-adults estimated using all 10 primaries was 2 September, 13 days later than for adults. The sub-adults’ primary moult was estimated to last on average 134 days, which did not differ signiicantly from the 131 days in adults. The rate primary feather mass is deposited did not differ between the sub-adults and the adults. Moult of P1 and P2 in sub-adults started 10–11 days later than in adults, but overlapped in the same manner as in adults. The number of primaries grown simultaneously with subsequently moulted primaries and the size of the wing gap in sub-adults resembled the pattern in adults. Sub-adults inished their primary moult on 15 January on average, 15 days later than adults. We suggest that sub-adult Wood Sandpipers catch up with the timing of the adults when they are 19–20 months old, when they inish their irst complete moult of primaries, before the pre-migratory fattening period in February–March. adults do (Ginn & Melville 1983, Pearson 1974, Prater 1981, Tree 1974). Fragmentary attempts have been made to describe this second-year moult of birds in several wader species, but these descriptions have been conined to the moult of birds spending a “gap year” in the southern hemisphere, in species such as Curlew Sandpipers Calidris ferruginea (Elliott et al. 1976, Underhill 2006, Waltner 1976), Ruddy Turnstones Arenaria interpres (Summers et al. 1989) and Grey Plovers Pluvialis squatarola (Serra et al. 1999). The course of this second-year complete moult has not been described in detail for any migrant wader species because of the paucity of suitable moult records and inadequate methods to analyse this moult. The development of the Underhill–Zucchini (1988) moult model and the realization that it could be applied to estimate moult parameters for single primaries (Serra 2000, Serra & Underhill 2006, Underhill 2003) has enabled studies of moult to be extended to include the partial moult of INTRODUCTION Migrant waders use a variety of moult patterns in their irst year, but usually replace all or some of the irst set of primaries on their non-breeding grounds. In small species these immatures usually moult all 10 primaries, but many mediumsize and most large waders replace a varying number of outer primaries (Ginn & Melville 1983, Pearson 1974, Prater 1981, Prater et al. 1977, Pyle 2008, Tree 1974). This moult is usually completed before the birds depart on their irst northward migration. Waders that take a “gap year” and spend the austral summer and the following winter on their non-breeding grounds, such as Curlew Sandpiper Calidris ferruginea, delay this moult for a few months (Elliott et al. 1976, Ginn & Melville 1983, Pyle 2008, Underhill 2006). The irst moult of primaries is usually followed by a complete replacement of all 10 primaries during the birds’ second year of life, as the 35 36 Wader Study Group Bulletin 117(1) 2010 this pattern with that of immature and adult birds to provide the full picture of the consecutive primary moult cycles in this species in their southernmost African non-breeding grounds. In particular we aim to determine if sub-adult birds’ irst complete primary moult differs in timing and duration from the following sequences of complete primary moult by adults. MATERIAL AND METHODS Fig. 1. Location of catching sites of Wood Sandpipers in southern Africa (black dots). Numbers in brackets indicate the number of birds trapped in Zimbabwe and in South Africa. immature waders (Remisiewicz et al. 2010). It has also enabled other patterns of moult to be investigated quantitatively. Wood Sandpipers Tringa glareola are migrants which breed from Fennoscandia to the Ural Mountains and have non-breeding grounds that in Africa extend from the subSahel region to southern Africa (Cramp & Simmons 1983, Piersma et al.1996, Scott 2009, Underhill et al. 1999). Eggs are laid from the end of April and in May, and are hatched in June and July (Cramp & Simmons 1983, Dementev & Gladkov 1951). So, we use June as the reference month for hatching and use the term immatures for Wood Sandpipers during their irst year of life, sub-adults – for birds during their second year of life, and adults – for unidentiied subadults or older birds. In contrast to other medium and larger waders that defer their irst northward migration until they are almost two years old and spend their irst austral winter in southern Africa (Summers et al. 1995, Underhill 2006), most Wood Sandpipers undertake their irst northward migration when they are about 10 months old (Cramp & Simmons 1983, Piersma et al. 1996, Remisiewicz et al. 2010, Underhill 1997). Earlier studies of the moult of Wood Sandpipers in their African non-breeding grounds provided general descriptions of the moult of immature and adult birds (Pearson 1974, Remisiewicz et al. 2010, Tree 1974). These studies showed that most immatures in southern Africa undertake a partial moult of two to six outer primaries and have a choice of strategies for this moult. Adults in southern Africa generally moult all ten primaries before migrating north (Remisiewicz et al. 2009). The moult of sub-adults remained a gap in our knowledge of the moult patterns of most migrant waders. We chose Wood Sandpipers for an analysis of this primary moult in migrant waders because in this species the northward migration by most immatures separates subsequent cycles of primary moult. This is in contrast with the pattern adopted by waders that take a “gap year”, among which the partial moult and the following complete moult of primaries might overlap (Underhill 2006). This paper aims to present the pattern of primary moult of sub-adult Wood Sandpipers in southern Africa and to compare We analysed the primary moult scores of 97 sub-adult Wood Sandpipers from two sources: 13 specimens from the National Museum of Zimbabwe in Bulawayo collected in 1900–1975 from Zimbabwe and Botswana, and 84 birds ringed in southern Africa in 1966–2008 during irregular mist-netting at wetlands in Zimbabwe and South Africa (Fig. 1). Wood Sandpipers were aged by their plumage (Prater et al. 1977) as immatures (birds 3–12 months old, from their arrival in August till 31 May of the following year), sub-adults (birds 13–20 months old, from 1 June of their second year of life to the end of their primary moult in the following December– January) and adults (unidentiied sub-adults and birds older than 20 months). Sub-adults, analysed in this paper, could be identiied either by the juvenile-type inner median coverts, which they might retain up to about 14 months (MR, unpubl. data) or by the different generations of primaries (Prater et al. 1977). Individuals analysed in this paper were identiied as sub-adults by the presence of one or both of these diagnostic features. The contrasting tones and wear between two generations of primaries can indicate completed partial moult of a few outer primaries, typical of immatures (Remisiewicz et al. 2010, Tree 1974). Some sub-adults replace the diagnostic coverts and all primaries earlier and then become indistinguishable from adults. All available moult records of sub-adults from all years and locations were combined, assuming them to be representative of the whole of southern Africa. The date of capture or collection was taken as the number of days from 1 June. The state of moult in the primaries of one wing was recorded as a moult formula, a string of 10 scores, one for each primary, following the standard approach of Ashmole (1962) and Ginn & Melville (1983). The sub-adults were divided into two main groups by these moult scores: birds in which contrasting patterns indicated an earlier partial moult of outer primaries and those that showed no traces of partial moult. In sub-adults that showed previous partial primary moult we treated both generations of fully grown primaries (the oldest juvenile primaries and the newer ones acquired in the partial moult) as the “old generation” feathers because both were grown before the complete moult started. We assigned both generations of the old primaries scores of 0. Primaries being grown in the complete moult were given scores of 1–5. Further calculations followed the methods used to analyse the primary moult of adult Wood Sandpipers (Remisiewicz et al. 2009). Based on the moult scores, and using the entire tract of all 10 primaries combined (minute outermost 11th primary ignored), we calculated for each individual the Proportion of Feather Mass Grown (PFMG) recommended by Underhill & Zucchini (1988). Assuming that primary feather tissue is deposited at a continuous rate, this provides the moult index required by the Underhill–Zucchini (1988) model. For this calculation we followed Underhill & Summers (1993) and used the mean relative masses of each primary for Wood Sandpipers given by Remisiewicz et al. (2009). We estimated three parameters of moult (mean starting date, its standard 37 Remisiewicz et al.: Primary moult in second-year Wood Sandpipers 2 6 33 30 16 9 1 Table 1. Percentages of outer primary feathers renewed in the partial moult during the irst year in sub-adult Wood Sandpipers in southern Africa, determined for immatures (n = 400, Remisiewicz et al. 2010) and sub-adults (n = 50). Calculations are restricted to birds which moulted outer primaries. No. of primaries replaced Primary feathers moulted 2 3 4 5 6 % among immatures P9–P10 P8–P10 P7–P10 P6–P10 P5–P10 % among sub-adults 1.3 6.5 63.3 27.3 1.8 0.0 2.0 55.1 38.8 4.1 Months Fig. 2. Percentages of sub-adult Wood Sandpipers caught in southern Africa with different moult status of primaries: white bars – birds showing partial moult of outer primaries and complete moult from P1 not yet started; dotted bars – birds showing partial moult and complete moult in progress; bars with diagonal lines – birds showing no partial moult and complete moult not yet started; black bars – birds showing no partial moult and complete moult in progress; numbers above bars show monthly sample sizes. deviation, and moult duration) using the Underhill–Zucchini moult models (Underhill & Zucchini 1988, Underhill et al. 1990) and applied the software designed to run these models (Brandão 1998, Underhill et al. 2006). We used the moult model for data of Type 5 of Underhill et al. (1990) because the sample contained only birds which had not yet started moult and birds in active moult; no birds which had completed moult were considered because they become indistinguishable from adults when they are close to completion of moult. The conidence limits for the moult starting date (during which 95% of birds are estimated to start moult) were calculated as the estimated mean starting date +1.96×standard deviation. Additionally we estimated the three moult parameters separately for individual primaries (Serra 2000, Underhill 2003), forming a moult index by transforming moult scores from 0 to 5 to the values 0, 0.125, 0.375, 0.625, 0.875 and 1 respectively. We applied the standard Underhill–Zucchini moult model using data Type 2 because the sample contained birds that had not yet started moult, birds in active moult and birds which had completed moult of the analysed primary. We estimated the size of the gap in the primary feathers during moult, accounting for the relative sizes of the missing feathers, by calculating Proportion of Feather Mass Missing (PFMM) for each individual (Remisiewicz et al. 2009, 2010, Ward et al. in press). Feathers that had moult scores 1, 2, 3 and 4 were estimated to be missing 0.875, 0.625, 0.375 and 0.125 respectively of the relative mass of the feathers. We compared the moult of sub-adult Wood Sandpipers with the results obtained for adults (Remisiewicz et al. 2009) and immatures (Remisiewicz et al. 2010) and made use of the data from those papers. We applied the G-test (Zar 1999) to compare proportions of immature and sub-adult birds showing different patterns of partial moult. We used the likelihood ratio test (Burnham & Anderson 1998) to test the null hypotheses that the starting dates of primary moult and of the moult of individual primaries for sub-adults were the same as for adults. RESULTS We classiied 97 Wood Sandpipers as sub-adults (Fig. 2). We recorded two in June and six in July, then 33 in August and 30 in September and fewer in subsequent months (Fig. 2). Three to six outer primaries had been replaced during the partial moult in 50 birds and 16 of these had already initiated a full moult from the innermost P1 (Fig. 2). Of the 50 sub-adults birds with partial moult, 94% had replaced either four or ive outer primaries (Table 2). Of the 47 sub-adults without partial moult, 38 were actively moulting, having started at P1 (Fig. 2). The earliest date for a sub-adult in active moult was 30 July. The 54 sub-adults in active moult had all started at P1 and were progressing without suspending to P10. From September the proportion in active moult gradually increased (Fig. 2). After December no Wood Sandpipers were identiied as sub-adults because they were indistinguishable from adults. To estimate the timing of moult of sub-adult Wood Sandpipers, we combined the samples showing a previous partial moult with those that had not moulted. This assumed that there were no biological reasons for these two groups to differ in the timing of their irst complete primary moult; the sample sizes were too small to verify this statistically. Table 2. Estimates of moult parameters of primaries P1 and P2 and of the 10 primaries as a single tract for sub-adults and adults (Remisiewicz et al. 2010) in Wood Sandpipers moulting in southern Africa. Primary and age group P1 adults P1 sub-adults P2 adults P2 sub-adults P1–P10 adults P1–P10 sub-adults Moult parameters Sample sizes Mean start date (SD) Duration (SD) Standard deviation of start date (SD) End date (SD) Not yet moulted In moult Moult complete Moult model used 22 Aug (1.6) 1 Sep (3.1) 23 Aug (1.5) 2 Sep (3.0) 21 Aug (1.7) 2 Sep (3.0) 25 (1.9) 27 (4.6) 25 (1.9) 27 (4.6) 131 (2.5) 134 (9.9) 19 (1.1) 19 (2.5) 19 (1.1) 18 (2.4) 29 (0.8) 20 (2.1) 16 Sep (1.5) 27 Sep (4.2) 17 Sep (1.1) 29 Sep (4.2) 30 Dec (1.4) 14 Jan (8.6) 108 42 114 44 108 53 109 23 108 22 687 42 1218 30 1213 29 640 0 Type 2 Type 2 Type 2 Type 2 Type 2 Type 5 38 Wader Study Group Bulletin 117(1) 2010 Number of shed primary Fig. 4. Simultaneously growing primaries of Wood Sandpipers in southern Africa: open squares – sub-adults; black circles – adults (Remisiewicz et al. 2009). Values shown are the means and 95% conidence intervals of the number of primaries growing while each primary is being replaced. Sample sizes for adults and sub-adults are shown below and above the lines representing the conidence intervals respectively. Fig. 3. Distribution through time of the Proportion of Feather Mass Grown (PFMG) of Wood Sandpipers of three age categories, corresponding with birds’ years of life: top igure – immatures (birds during their irst year of life, n = 674) (Remisiewicz et al. 2010); middle igure – sub-adults (birds during their second year of life, n = 97), bottom graph – adults (n = 1495) (Remisiewicz et al. 2009). The parallelograms show the temporal limits of moult for 95% of birds; the thick lines through the centre of the parallelograms show the estimated timing of moult; in the top igure: solid lines – immatures that moult six primaries in partial moult, dashed lines – immatures that moult three primaries in partial moult. The PFMG values of two individuals were identiied by the moult model as outliers: formula 5555555322 on 27 July and 5555555315 on 7 September. For the remaining 95 sub-adults the mean starting date of the primary moult, estimated using the whole primary feather tract, was 2 September (Table 1, Fig. 3) with the standard deviation of the starting date estimated to be 20 days (Table 2), so that 95% were estimated to begin moult during the 77-day period 26 July–11 October (Fig. 3). The estimated duration of the primary moult was 134 days so that the mean completion date was 14 January and 95% of sub-adults completed primary moult in the period 7 December–21 February (Table 2, Fig. 3). From the duration of moult (134 days) we estimated that the rate of production of the primary feather material, if uniform, was 0.75% PFMG per day. For individual primaries, the mean starting date of moult of P1 was estimated to be 1 September and that of P2 was one day later; the moult duration for both primaries was 27 days (Table 2). The numbers of birds in active moult of other primaries was too small to obtain estimates of those moult parameters. The number of primaries growing simultaneously varied from one to four (Fig. 4). During the replacement of P1–P5 the mean number of feathers growing simultaneously was larger than three and it decreased to less than two during the replacement of P6–P8; for P9 and P10 the samples were inadequate for analysis. The average size of the wing gap, expressed as the Proportion of Feather Mass Missing (PFMM), in sub-adults in active moult was 0.11 (SD = 0.052, n = 44, range 0.01–0.25) (Fig. 5). On average 11% of the mass of primary light feathers was missing during moult. In all except two sub-adults less than 19% was missing; the largest values were 25% in a bird replacing its three outer primaries simultaneously (formula 5555555322 on 27 July) 24% in a bird replacing ive inner feathers simultaneously (formula 2221100000 on 23 September). In sub-adults in active moult, the PFMM was not signiicantly correlated with PFMG (Fig. 5; r = –0.23, n = 52, p = 0.13). DISCUSSION Timing of moult of sub-adult Wood Sandpipers in southern Africa The small numbers of sub-adult Wood Sandpipers recorded in June and July in southern Africa were most likely the birds that did not migrate northwards at the end of their irst year of life. This is in agreement with observations of a few immature Wood Sandpipers in southern Africa in May, after the species’ main departure time between mid-March and mid-April (Herremans 1994, Irwin 1981, Remisiewicz et al. 2010, Tarboton et al. 1987, Taylor 1979, Underhill et al. 1997); in June–July these birds enter their second year of life i.e. become sub-adults. The increasing numbers trapped in August and September relect the arrival of migrant sub- Remisiewicz et al.: Primary moult in second-year Wood Sandpipers 39 adults, at the same time as adults arrive in southern Africa (Underhill 1997). Comparison of moult between sub-adults and immatures The most frequent pattern of partial moult in immatures is the replacement of the outer four or ive primaries; 94% of immature moulters adopt this strategy (Remisiewicz et al. 2010). We observed the same pattern in sub-adults (Table 1). The proportions of immatures and sub-adults moulting various numbers of primaries were not signiicantly different (Table 1, G-test: G3 = 5.97, p = 0.11). This is what we would expect, considering that in July the immatures that stayed in southern Africa turn into sub-adults. The small sample of sub-adults caught in June and July did not allow us to check if any particular strategy of partial moult is preferred by birds of this age group staying in southern Africa. Comparison of moult between sub-adults and adults Fig. 5. The relationship between the size of the gap in the primaries during moult, expressed as Proportion of Feather Mass Missing (PFMM) and Proportion of Feather Mass Grown (PFMG) for sub-adult Wood Sandpipers in active moult in southern Africa. Based on the whole primary feather tract, sub-adults started moult 13 days later on average than adults (Table 2, Fig. 2). This difference was signiicant; we performed a likelihood ratio test by irst pooling the two data sets and estimating a single set of parameters and then allowing the sub-adults and adults to have different starting dates, and assuming a common duration and SD of starting date (χ21 = 21.73; p < 0.001). The mean moult starting date estimated for P1 was 10 days later in sub-adults than in adults (Table 2); and the similar likelihood ratio test showed that this difference was also signiicant (χ21 = 13.01; p < 0.001). The moult starting date estimated for P2 in sub-adults was 11 days later than for adults (Table 2); this difference was again signiicant (χ21 = 14.39; p < 0.001). The duration of moult for all primaries combined was estimated at 134 days in sub-adults and at 131 days in adults, but this difference was not signiicant (z = 1.00; p = 0.16). The estimated date of completion of moult by sub-adults was 14 January, 15 days later than the adults (Table 2, Fig. 2). But the two-week delay in completing the primary moult in southern Africa should not affect sub-adults’ ability to catch up with adults as they prepare for their next departure to the breeding grounds. Adult Wood Sandpipers begin fattening in mid-February (authors’ unpubl. data) before leaving between mid-March and midApril. Thus the sub-adults would still have at least a month from the end of their primary moult to the beginning of the fattening period; during this time they could also inish moulting their secondaries, tertials, rectrices and advance with body moult into breeding plumage (authors’ unpubl. data). But the northwards passage of sub-adults through stopover sites close to the breeding grounds was on average three days later than of adults (Remisiewicz & Wennerberg 2006). This might relect slightly later departures of sub-adults than adults for the breeding grounds, but also slower migration of less experienced sub-adult birds, similar to that described during their irst southward migration (Wichmann et al. 2004). Sub-adults and adults showed a similar pattern in the number of primaries grown simultaneously during the replacement of each primary (Fig. 4). The mean values during the replacement of each of the primaries P1–P5 did not differ significantly between sub-adults and adults (t-test, for each primary: p > 0.05); the samples for the remaining primaries in sub-adults were too small to compare. The gap in sub-adults’ wings caused by shedding primaries averaged 11% of the mass of primaries, similar to that of adults (10%) (Remisiewicz et al. 2009) and of immatures (8–11% in different groups of moulters) (Remisiewicz et al. 2010). These values did not differ signiicantly between the three age groups (Kruskal–Wallis test: H2, 1031 = 1.28, p = 0.53). In adult Wood Sandpipers PFMM was signiicantly correlated with PFMG (r = –0.19, p < 0.001), but there was no similar relationship in sub-adults. But in both age classes the trend shown by the correlation coeficient was negative, so that gaps tended to be smaller at the start of moult than at the end. These relatively small wing gaps should not substantially impede the immature Wood Sandpipers’ light ability and would allow them to move between ephemeral wetlands during their moult in southern Africa (Remisiewicz et al. 2009). Comparison with the moult patterns in sub-adults of other wader species Migrant waders that spend a “gap year” in the southern hemisphere usually moult all their primaries in their second year soon after undertaking a partial moult of their primaries. These two moult sequences sometimes overlap (Pyle 2008). The overlap of two consecutive cycles of moult has been described in sub-adult Curlew Sandpipers Calidris ferruginea (Elliott et al. 1976, Underhill 2006, Waltner 1976), Ruddy Turnstones Arenaria interpres (Summers et al. 1989) and Grey Plovers Pluvialis squatarola (Serra et al. 1999). These waders spend their first austral winter in southern Africa (Summers et al. 1995). Overlapping moult has never been observed in Wood Sandpipers in southern Africa, probably because these irst two moults of primaries are separated by migration to and from the breeding grounds in most birds (Remisiewicz et al. 2010). In Curlew Sandpipers, which do spend a “gap year” in southern Africa or Australia, the question of whether the duration and timing of the complete moult of primaries in subadults differs from that of adults remains unanswered (Elliott et al. 1976, Minton et al. 2006, Underhill 2006). Summers 40 Wader Study Group Bulletin 117(1) 2010 et al. (1989) described the second cycle of primary moult in Ruddy Turnstones in southern Africa as so slow that some second-year birds had not inished this moult by the time of their northwards departure. This was not the case in Wood Sandpipers, in which the rate of the second moult of primaries in sub-adults was the same as in adults, and this moult was completed about two months before the departure period. CONCLUSIONS Our results showed that the complete moult of primaries of sub-adult Wood Sandpipers is similar to that of adults in duration, rate of primary feather mass deposition, the number of primaries grown simultaneously and the size of the gap in the wing. But the moult of sub-adults starts and ends about two weeks later than in the adults. We suggest that the sub-adult Wood Sandpipers catch up with the timing of the adults after they complete this primary moult, when they are 19–20 months old, during the pre-migratory fattening period in February–March, before they depart from southern Africa between mid-March and mid-April. ACKNOWLEDGEMENTS Financial support to MR and LGU was provided by the National Research Foundation (NRF), South Africa, and the University of Gdańsk, Poland, through the Poland– South Africa Agreement in Science and Technology. Anna Gustowska and Włodek Meissner assisted with ieldwork in KwaZulu-Natal. MR was supported by a postdoctoral fellowship from the Claude Leon Foundation. LGU acknowledges support from the SeaChange Programme of the NRF. We thank David Melville for reviewing an early draft of our paper. Joel Avni assisted with language correction and editing. REFERENCES Ashmole, N.P. 1962. The Black Noddy, Anous tenuirostris, on Ascension Island. Part 1. General biology. Ibis 103b: 235–273. Brandão, A. 1998. A comparative study of stochastic models in biology. Unpubl. Ph.D. thesis, University of Cape Town, Cape Town. Burnham, K.P. & Anderson, D.R. 2002. Model selection and multimodel inference: a practical information–theoretic approach. 2nd ed. SpringerVerlag, New York. p. 337. Cramp, S. & Simmons, K.E.L. eds. 1983. The Birds of the Western Palearctic. Vol. 3. Oxford University Press, Oxford. Dementiev, G.P. & Gladkov, H.A. eds. 1951. Birds of USSR. Vol. 3. Nauka, Moscow. [in Russian]. Elliott, C.C.H., Waltner, M., Underhill, L.G., Pringle, J.S. & Dick, W.J.A. 1976. The migration system of the Curlew Sandpiper Calidris ferruginea in Africa. Ostrich 47: 191–213. Ginn, H.B. & Melville, D.S. 1983. Moult in Birds. BTO Guide 19. British Trust for Ornithology, Tring. Herremans, M. 1994. Fifteen years of migrant phenology records in Botswana: a summary and prospects. Babbler 28: 47–68. Irwin, M.P.S. 1981. The Birds of Zimbabwe. Quest, Salisbury. Minton, C.D.T., Rogers, K.G., Jessop, R.E., Graham, D.M. & Lowther, A.D. 2006. Biometrics and moult of the Curlew Sandpiper Calidris ferruginea in Australia. Int. Wader Stud. 19: 195–204. Newton, I. 2008. The Migration Ecology of Birds. Academic Press, London. Pearson, D. 1974. The timing of wing moult in some Palaearctic waders wintering in East Africa. Wader Study Group Bull. 12: 10–17. Piersma, T., van Gils. J. & Wiersma P. 1996. Family Scolopaciade (sandpipers, snipes and phalaropes). In: del Hoyo, J., Elliot, A. & Sargatal, J. (eds). Handbook of the Birds of the World. Vol. 3. Hoatzin to Auks. Lynx Edicions, Barcelona. pp. 444–533. Prater, A.J. 1981. A review of the patterns of primary moult in Palearctic waders (Charadrii). In: Cooper, J. (ed.). Proceedings of the Symposium on Birds of the Sea and Shore. African Seabird Group, Cape Town. pp. 393–409. Prater, A.J., Marchant, J.H. & Vuorinen, J. 1977. Guide to the Identiication and Ageing of Holarctic Waders. BTO Guide 17. British Trust for Ornithology, Tring. Pyle, P. 2008. Identiication Guide to North American Birds, Part 2. Slate Creek Press, Bolinas, California. pp. 500–507. Remisiewicz, M. & Wennerberg, L. 2006. Differential migration strategies of Wood Sandpiper (Tringa glareola) – genetic analyses reveal sex differences in morphology and spring migration phenology. Ornis Fennica 83: 1–10. Remisiewicz, M., Tree, A.J., Underhill, L.G., Gustowska, A. & Taylor, P.B. 2009. Extended primary moult as an adaptation of adult Wood Sandpipers Tringa glareola to their use of freshwater habitats of southern Africa. Ardea 97 (3): 271–280. Remisiewicz, M., Tree, A.J., Underhill, L.G. & Taylor, P.B. 2010. Rapid or slow moult? The choice of a primary moult strategy by immature Wood Sandpipers Tringa glareola in southern Africa. J. Ornithol. 151: 429–441. Scott, D. 2009. Wood Sandpiper Tringa glareola. In: An atlas of wader populations in Africa and western Eurasia. Delany, S., Scott, D., Dodman, T., Stroud, D. (eds), Davidson, N., Kirby, J., Underhill, L.G. (assistant eds). Wageningen, Wetlands International. pp. 338–342. Serra, L. 2000. How do Palearctic Grey Plovers adapt primary moult to time constraints? An overview across four continents. Wader Study Group Bull. 93: 11–12. Serra, L., Whitelaw, D.A., Tree, A.J. & Underhill, L.G. 1999. Moult, mass and migration of Grey Plovers Pluvialis squatarola wintering in South Africa. Ardea 87: 71–81. Serra, L. & Underhill, L.G. 2006. The regulation of primary moult speed in the Grey Plover, Pluvialis squatarola. Acta Zool. Sin. 52 (Suppl.): 451–455. Summers, R.W., Underhill, L.G., Clinning, C.F. & Nicoll, M. 1989. Populations, migrations, biometrics and moult of the Turnstone Arenaria i. interpres on the East Atlantic coastline, with special reference to the Siberian population. Ardea 77: 145–168. Summers, R.W., Underhill, L.G. & Prŷs-Jones, R.P. 1995. Why do young waders in southern Africa delay their irst return migration to the breeding grounds? Ardea 83: 351–357. Tarboton, W.R., Kemp, M.I. & Kemp, A.C. 1987. Birds of the Transvaal. Transvaal Museum, Pretoria. Taylor, P.B. 1979. Palaearctic and intra-African migrant birds in Zambia: A report for the period May 1971 to December 1976. Zambian Ornithological Society Occasional Paper 1. Tree, A.J. 1974. The use of primary moult in ageing the 6–16 month age class of some Palearctic waders. Safring News 3(3): 21–24. Underhill, L.G. 1997. Wood Sandpiper Tringa glareola. In: Harrison, J.A., Allan, D.G., Underhill, L.G., Herremans, M., Tree, A.J., Parker, V. & Brown, C.J. (eds). The Atlas of Southern African Birds. Vol. 1: Nonpasserines. BirdLife South Africa, Johannesburg, pp. 410–411. Underhill, L.G. 2003. Within ten feathers: Primary moult strategies of migratory waders (Charadrii). In: Avian Migration. Berthold, P., Gwinner, E. & Sonnenschein, E. (eds). Springer-Verlag, Berlin Heidelberg, pp. 187–197. Underhill, L.G. 2006. A preliminary overview of the life spiral of Curlew Sandpipers Calidris ferruginea. Int. Wader Stud. 19: 209–211. Underhill, L.G. & Summers, R.W. 1993. Relative masses of primary feathers in waders. Wader Study Group Bull. 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 Redshank Tringa totanus. Ibis 132: 118–123. Underhill, L.G., Tree, A.J., Oschadleus, H.D. & Parker, V. 1999. Review of ringing recoveries of waterbirds in southern Africa. Avian Demography Unit, University of Cape Town, Cape Town. Underhill, L.G., Serra, L. & Brandão, A. 2006. Progress with the statistical analysis of primary molt. Acta Zool. Sin. 52 (Supplement): 440–443. Waltner, M. 1976. Moult in Palaearctic waders. Safring News 5: 14–16. Ward, V.L., Oschadleus, H.D. & Underhill, L.G. in press. Primary moult of the Kelp Gull Larus dominicanus vetula in the Western Cape, South Africa. Acta Orn. Wichmann, G., Barker, J., Zuna-Kratky, T., Donnerbaum, K.& Rössler, M 2004. Age-related stopover strategies in the Wood Sandpiper Tringa glareola. Ornis Fennica 81: 169–179. Zar, J.H. 1999. Biostatistical Analyses. Prentice-Hall, Inc., Upper Saddle River, New Jersey. .