Sex differences in the stopover ecology of Curlew Sandpipers Calidris ferruginea at a refuelling area during autumn migration JORDI FIGUEROLA* and ALBERT BERTOLERO Department of Animal Biology (Vertebrate), Faculty of Biology, University of Barcelona, Avda. Diagonal 645, E-08028 Barcelona, Spain We investigated the stopover patterns of male and female Curlew Sandpipers at a stopover area in northeast Spain. Curlew Sandpipers were trapped and colour-ringed during autumn migration in 1992 and 1993. Stopover length was similar to those reported previously for this and other waders that migrate using a small number of widely separated staging areas, but were greater than stopovers reported for other waders that migrate using a large number of staging areas separated by short distances. The differences in stopover length between the birds using these two strategies could be related to the fuel reserves that have to be accumulated to reach the next staging area. Males stayed longer in the area than females. Seasonal changes in prey availability or sex differences in moulting and migratory patterns do not account for these differences in stopover ecology. Following a time-selected model of optimal migration, sex differences in stopover ecology could be related to a dominance of the larger females over the males or to a higher foraging efficiency or a shorter search and settling time in females. Whether these differences are restricted to the studied area or are widespread in other staging areas used by the species could be important for assessing the possible differences in the migration speed of Curlew Sandpipers. The finding that males leave the breeding grounds 21—35 days before females but arrive at the study area with only a 10-day difference supports the hypothesis that females migrate faster than males at least in the first half of their migration. D uring migration, long-distance migrants have to replenish their energetic reserves several times at a number of staging areas.1 The overall time needed to complete migration is limited by the time necessary to accumulate new reserves2 (short periods of flight alternate with longer ones of fat accumulation) and, to a lesser extent, by flight speed. Two strategies of migration have been described and documented mainly in waders.3 The first strategy is used by ‘hoppers’ which make short flights and use a large number of staging areas along their *Correspondence author. Email: jordif@porthos.bio.ub.es migratory route. The second is used by ‘jumpers’ which fly large distances between the small number of staging areas they use to complete their migration. Alerstam & Lindström2 produced a model based on the optimization criteria of the decisions faced by a bird accumulating fat at a staging area. These models assumed that a bird assessed the moment to leave a staging area in terms of its present fat reserves, rate of fat accumulation and its expected future fat gain.2,4 These authors presented different models to explain the decisions of a bird at a stopover area. The differences between these models rely on the variables the birds try to optimize. ‘Time-selected’ migrants are expected to maximize migration speed and so are very sensitive to variations in their fat deposition rate. According to this model, birds with higher intake rates or which select habitats with greater food resources will migrate faster than other birds that do not forage as efficiently or use less favourable habitats. ‘Energy-selected’ migrants try to complete their migration with lower energetic costs by maximizing the relationship between flight range and fat load. The utility of accumulating additional fat decreases with increasing fat reserves because heavy body mass increases flight costs.5 Alerstam & Lindström’s model predicts that migration decisions of energyselected migrants are less sensitive to variations in fat deposition rate. According to the time-selected migration model, dominant individuals or more efficient foragers should depart earlier and/or with greater fat reserves than subdominants because dominants will be able to forage in more favourable areas. In species with energyselected migration, dominants should depart with lower fat loads than subdominants and, according to this model, no difference in stopover length is to be expected between these two groups. In this model the primary benefit for dominants is a reduced cost for search and settling.2 The effect of social status on departure mass and stopover length in time-selected migrants could explain the differences in stopover ecology of young and adults of several species during autumn migration6,7 and of males and females in spring.8,9 Despite the fact that most studies of the stopover ecology of waders have found agerelated differences in the residence patterns in autumn,7,1O no study has focused on the stopover patterns of males and females at a staging area during the autumn migration. Such studies could provide basic information to determine the possible intraspecific differences in the costs of migration, information necessary to determine the factors responsible for the winter distribution of waders.11 In this study we analyse the stopover patterns of adult male and female Curlew Sandpipers Calidris ferruginea at a staging area in autumn. The stopover ecology of this species was examined in relation to the optimization models of Alerstam & Lindström2 to test the hypothesis that the sexes differ in their stopover length. METHODS Curlew Sandpiper migration was studied at Les Salines de la Trinitat (4O°37′N OO°35′E; northeast Spain). The study area consisted of an area of mud and sand-flats in a saltpan near the sandy coast of the Mediterranean Sea. Little tidal oscillation occurs on this coast and the greatest variation in water level in the saltpan was produced by changes in wind force and direction, or when salt extraction was in progress. Birds were trapped at night by mist-net and by day using walk-in-traps from 27 July to 12 October 1992 and 25 July to 24 September 1993. Birds were colour-ringed and body mass, wing, bill and tarsus lengths of most birds were measured.12 Individuals were aged12 and the sex of adult birds (more than one-year-old) was determined from wing and bill length using the discriminant formula of Wymenga et al.13 Body mass at ringing was corrected to remove the effect of water loss and the emptying of the gut occurring between capture and data recording, assuming a rate of mass loss of O.88% per hour of captivity.14 Data from moulting birds were not included in the analysis because moulting increases the energetic requirements of birds15,16 and the rate of reserve accumulation at the study area was lower in moulting than in non-moulting Curlew Sandpipers. 17 Because the rate of reserve accumulation was one of the variables that affect stopover decisions,2 the inclusion of these birds would have affected the results of our analysis. Estimation of stopover length Approximately every other day a transect of 2.2 km was surveyed to locate and identify colourringed birds present in the area. From this transect we censused 44 ha of mud and sandflats, covering most of the suitable habitat for Curlew Sandpipers in the saltpan. Additional records of colour-marked birds obtained from casual observations were also considered. Estimating true stopover length is difficult because birds were usually not trapped immediately after arrival nor just before leaving the study area. To estimate stopover length we used the method described in Holmgren et al.7 This method is based on a mark—recapture model derived from Jolly18 and Seber.19 The assumptions of this model are that each individual has a constant probability of staying in the area and that the probability of resighting was the same for all birds. Using this model we calculated the maximum likelihood estimate of the stopover probability (ϖ) and the estimate of mean stopover length (τ) separately for each sex on each year (see Holmgren et al.7 for a detailed description of the model and parameter estimation). Statistical analyses Differences in stopover probabilities were tested by calculating the standard normal deviate (d) and the associated P value.7 Differences in body condition among groups were tested with square-root-transformed body mass using a trifactorial ANCOVA. In this analysis sex (male/female), control (resighted/not resighted) and year (1992/1993) were used as factors. Culmen length was introduced as a covariate to remove the effect of body size on body mass. When significant interactions between the factors were detected we used Scheffè multiple range tests to examine the differences between particular levels of the factors.2O RESULTS Relationship of body mass and stopover length Although females were larger than males,21 differences in body mass were not significant after controlling for the effect of bird size (three-way ANCOVA, F1,9O1 = O.27, P = O.61) and none of the interactions of sex with the other two factors was significant (sex—control: F1,9O1 = 2.11, P = O.15; sex—year: F1,9O1 = 1.27, P = O.26). Body mass of marked birds was greater in 1992 than in 1993 (three-way ANCOVA, F1,9O1 = 17.O7, P < O.OOO1). Body mass of birds resighted did not differ from those that were not resighted (three-way ANCOVA, factor control, F1,9O1 = 1.27, P = O.26). However, a significant interaction was detected between two of the factors (control— year, F1,9O1 = 4.29, P = O.O4). This significant interaction appeared because birds resighted in 1993 were leaner than birds trapped in 1992 (both resighted and not resighted, Scheffè tests, P < O.OO3 for both comparisons and P > O.2O for the rest of contrast). The relationships between body mass at ringing and stopover length were also analysed separately for each sex and year. None of the linear regressions between body mass and stopover lengthwas significant (r ranging from —O.O5 to —O.12 and P from O.3O to O.71). Frequency and length of stopover In 1992 we colour-marked 517 individuals, of which 99 were later resighted (19.1%). In 1993, 89 out of 393 marked individuals were resighted after the date of ringing (22.6%). Numbers of marked and resighted individuals did not differ between years (χ2 = 1.67, 1 df, P = O.2O), nor between sexes (1992: χ2 = 1.43, 1 df, P = O.23; 1993: χ2 = O.45, 1 df, P = O.5O). The estimated stopover length of males was longer than females in 1992 (d = 1.84, P < O.O5), in 1993 (d = 1.87, P < O.O5), and when data of both years were combined (d = 2.64, P < O.OO5), see Table 1. Table 1. Observed mean stopover length of males and females resighted at least once (t) and estimated mean stopover length of the Curlew Sandpiper population (τ) both expressed in days. Daily stopover probability (ϖ) and its standard deviation. Males Females Year t n τ ϖ sd t n τ ϖ sd 1992 15.O 56 13.O O.93 O.OO9 1O.8 49 8.8 O.9O O.O1O 1993 16.8 5O 14.O O.93 O.OO9 12.O 44 9.2 O.9O O.O1O Total 15.9 1O6 13.5 O.93 O.OO6 11.4 93 9.O O.9O O.O1O DISCUSSION Stopover length and migration strategy Comparing the stopover lengths reported in different studies is difficult because authors usually use different methods to estimate mean stopover length. Accounting for this shortcoming, the stopover lengths presented in this study were similar to the approximately 14 days reported at a staging area on the German Waddensea in autumn 1992,22 and fall within the range of 7—21 days for individuals refuelling on a British estuary.23 Similar stopover lengths were also reported for the Semipalmated Sandpiper Calidris pusilla in North America.24,25 Both species were considered to use a jumping strategy.24,26 However, all these estimates were higher than the stopover lengths reported for other waders of similar size that were considered to use a hopping strategy.1O,27,28 All Dunlins Calidris alpina migrating through Ottenby stayed in the area for less than nine days (a mean of O.55—1.44 days7) and very short stopovers were also reported for the Western Sandpiper Calidris mauri and Least Sandpiper Calidris minutilla during autumn migration along the western coast of North America.27,29 The differences in flight distance between species that jump and those that hop suggest that jumpers have to depart with higher levels of reserves and have to stay longer in the staging areas to replenish their fat reserves. Although theoretical flight ranges of jumpers and hoppers ready to depart in spring did not differ at a wintering area in Mauritania,14 the comparison of stopover lengths of several species in different areas suggests that jumpers will stop for longer periods in the staging areas to replenish their fat reserves. The apparent discrepancy between these two results could occur because both jumping and hopping strategies might occur in the same species and thus among the individuals using the same refuelling area. This pattern has been demonstrated by Skagen and Knopf3O with the use of radiotracking to describe the residency patterns of Semi-palmated Sandpipers and White-rumped Sandpipers Calidris fuscicollis at a staging area in spring. Their study showed that residency patterns were intraspecifically variable and although some birds departed with large fat loads and in a position to carry out long uninterrupted flights, others leave with small reserves and were only capable of short hops. The occurrence in one species of both migration strategies could explain why differences in flight range were not found by Zwarts et al.14 and that reported estimates of stopover length (a variable probably more influenced by birds staying longer in the area) were greater in jumpers than in hoppers. Optimal models of bird migration predict that in time-selected migrants, leaner birds will make longer stopovers than fatter birds. We did not find this relationship in our study nor in other studies of several species of waders with time-selected migration7,25,29,31,32 but see Mascher.33 As Holmgren et al.7 pointed out, these apparently random stopovers could appear if birds tend to wait for suitable weather conditions in order to resume migration and if weather changes randomly. In this situation time-selectedmigrants would continue to deposit fat, but energy-selected migrants would not. Sex differences in stopover ecology Our results show that, in autumn, male Curlew Sandpipers stayed for longer than females at the studied stopover area. Several factors could have caused these differences. Male Curlew Sandpipers migrate ahead of the females which remain in the breeding grounds taking care of the brood.21,26 The density of prey available for birds at a staging area could change during autumn migration and prey depletion by early migrants has been reported in several studies (reviewed in Szèkely & Bamberger34). Males arrived at our study area ten days before females in 1993 and prey density available for each sex could have changed in this time. However, in 1992 the breeding success of arctic waders was very low35 and in this year probably led to males arriving only four days before females.36 If temporal variation in prey availability was the cause of the differences in stopover ecology reported in this study, differences in stopover length of males and females should appear in 1993 but not in 1992 (when both sexes migrated approximately at the same time and stayed in the study area under the same biotic and abiotic conditions). This, however, is not the case and so we reject the hypothesis that temporal variation in prey availability may have produced the differences in stopover length reported in this study. Another possible source of bias in our estimates of stopover length could be related to the inclusion in the data set of birds in active wing moult. Although we have excluded from our analysis birds trapped in active wing moult, it is possible that some birds started primary moult just after being ringed, thereby increasing our estimates of stopover length. Some studies have reported that male Curlew Sandpipers start to moult before females,26,37,38 although others failed to find any difference.39,4O More males starting primary moult in our study area could have increased the estimated stopover length of this sex. However, the number of birds that started wing moult in the study area was small (6.2% of the adult birds ringed in autumns 1992 and 1993), and although more males started to moult in the study area in 1993, no difference in the number of males and females moulting in the area was detected in 1992.17 Longer stopovers of males were reported in both years, so we also reject the possibility that differences in moult patterns were the cause of the reported sex differences in stopover length (at least of those found in 1992). According to the models of optimal migration,2 a variety of factors could explain the sex differences in stopover length. Females could stay for a shorter period in the study area if they leave with lower levels of accumulated reserves. To examine this issue, Figuerola & Bertolero41 compared theoretical flight ranges of males and females staying at the Ebro Delta. Arrival body mass was calculated as the average body mass of the 1O% of leanest birds trapped in the area, and departure body mass was calculated as the average body mass of the 1O% of heaviest birds. These approximate calculations suggested that males accumulated approximately 3O.4 g and females 32.5 g of body reserves. If these figures reflect the real situation, the sexes did not differ greatly in their estimated flight range (3171 versus 3288 km) and the differences were not consistent with the hypothesis that females leave with smaller fat loads than males. The time-selected model of optimal migration suggests that birds with higher fat deposition rates or a lower search-settling time (time between arrival at the area and initiation of efficient fat deposition) could depart earlier from the staging areas.2 These differences in fat deposition rate and search costs could appear if one sex dominates the other or if sexes use ecologically different resources. In a field experiment at a staging area in Scandinavia, Lindström et al.42 demonstrated that dominant Bluethroats Luscinia svecica accumulated fat at a higher rate than subdominant ones. Although information on the dominance structure of Curlew Sandpiper populations is scarce (but see Puttick43), it is common among birds that the larger sex dominates the smaller one.44,45 Female Curlew Sandpipers are larger than males so it is fair to expect that they dominate males. Sexual dimorphism in bill length is also acute and could allow females to capture prey buried more deeply in the mud. According to these differences Puttick46 found in a wintering area in South Africa that the diet of the sexes differed and that females dominated males, foraged more successfully and needed less time foraging to cover their energetic requirements.43 All these observations lend support to the hypothesis that the sex differences in stopover length reported in this study could be produced by differences between sexes in the capacity to accumulate fat in the study area and/or because females can start to deposit fat earlier than males (i.e. have a shorter searchsettling time). Our study area could be especially suitable for females, whereas other areas could be more favourable for males. If the differences we found in the stopover ecology of Curlew Sandpiper also occur in the other staging areas used by the species, this would mean that the females had a higher overall migration speed, because this sex will stay at the stopover areas for less time than males. Males leave breeding grounds over three weeks before females21 but arrived at our study area in 1993 only 1O days ahead of females.36 This reduction of the differences in the timing of passage for each sex gives support to the hypothesis that sexes differ in their migration speed. Obviously, the results presented here do not demonstrate that in Curlew Sandpipers sexes differ in their migration efficiency. However, they are consistent with this hypothesis. We hope that this paper might stimulate further research on sex differences in stopover ecology and searchsettling patterns in species with significant normal or reverse sexual dimorphism in body size. ACKNOWLEDGEMENTS This work was funded by the Ebro Delta Natural Park, Diputació de Tarragona and Grup Català d ‘Anellament. The Generalitat de Catalunya (Catalan Autonomous Government) provided the permits necessary for carrying out the fieldwork. L.M. Copete and L. Gustamante made an indispensable contribution to the fieldwork. We would also like to thank J.M. Arcos, L. Brotons, L. Carrera, D. Escobar (Fonoteca del Museo de Zoologia de Barcelona), M.A. Franch, D. Froelich, S. Galan, R. Marine, R. Marti, A. Martinez, R. Mateo, A. Motis, D. Oró, X. Riera, A. Salmerón, J. Solans, F. Vicents and R. Vidal for their help and friendship. M. Lockwood improved the English and J.A. Amat, B.J. Ens, J.D. Goss-Custard, A. Lindström, T. Piersma, J.C. Senar and D.A. Stroud made valuable comments on earlier drafts of this manuscript. REFERENCES 1. Alerstam, T. (199O) Bird Migration. Cambridge University Press, Cambridge. 2. Alerstam, T. & Lindström, A. (199O) Optimal bird migration: the relative importance of time, energy and safety. In Bird Migration: Physiology and Ecophysiology (ed. E. Gwinner), pp. 331—351. Springer, Berlin. 3. Piersma, T. (1987) Hop, skip or jump? Constraints on migration of arctic waders by feeding, fattening and flight speed. Limosa, 60, 185—194. 4. Lindström, A. (199O) Stopover ecology of migrating birds. PhD Thesis, Department of Ecology, Lund University. 5. Pennycuick, C.J. (1975) Mechanics of flight. In Avian Biology 5 (eds D.S. Farner & J.R. King), pp. 1—75. Academic Press, New York. 6. Ellegren, H. (1991) Stopover ecology of autumn migrating Bluethroats Luscinia s. svecica in relation to age and sex. Ornis Scand., 22, 34O—348. 7. Holmgren, N., Ellegren, H. & Pettersson, J. (1993) Stopover length, body mass and fuel deposition rate in autumn migrating adult Dunlins Calidris alpina: evaluating the effects of moulting status and age. Ardea, 81, 9—2O. 8. Lavee, D., Safriel, U.N. & Meilijson, I. (1991) For how long do trans-Saharan migrants stop over at an oasis? Ornis Scand., 22, 33—44. 9. Morris, S.R., Richmond, M.E. & Holmes, D.W. (1994) Patterns of stopover by warblers during spring and fall migration on Appledore Island, Maine. Wilson Bull., 106, 7O3—718. 1O. Pienkowski, M.W. & Dick, W.J.A. (1975) The migration and wintering of Dunlin Calidris alpina in northwest Africa. Ornis Scand., 6, 151—167. 11. Meltofte, H. (1996) Are african wintering waders really forced south by competition from northerly wintering conspecifics? Benefits and constraints of northern versus southern wintering and breeding in waders. Ardea, 84, 31—44. 12. Prater, A.J., Marchant, J.H. & Vourinen, J. (1977) Guide to the Identification and Ageing of Holarctic Waders. BTO, Tring. 13. Wymenga, E., Engelmoer, M., Smit, C.J. & van Spanje, T.M. (199O) Geographical origin and migration of waders wintering in West Africa. Ardea, 78, 83—112. 14. Zwarts, L., Ens, B.J., Kersten, M. & Piersma, T. (199O) Moult, mass and flight range of waders ready to take off for long-distance migrations. Ardea, 78, 339—364. 15. Payne, R.B. (1972) Mechanisms and control of moult. In Avian Biology 2 (eds. D.S. Farner & J.R. King), pp. 1O3—155. Academic Press, New York. 16. Lindström, A., Visser, G.H. & Daan, S. (1993) The energetic cost of feather synthesis is proportional to basal metabolic rate. Physiol. Zool., 66, 49O—51O. 17. Figuerola, J. & Bertolero, A. (1995) The primary moult of Curlew Sandpiper in the Ebro Delta, North-East Spain. Ringing Migr., 16, 168—171. 18. Jolly, G.M. (1965) Explicit estimates from capturerecapture data with both death and immigration-stochastic model. Biometrika, 52, 225—247. 19. Seber, G.A.F. (1973) The Estimation of Animal Abundance and Related Parameters. Macmillan, London. 2O. Cuadras, C.M. (1991) Problemas de Probabilidades y EstadIstica. Vol. 2: Inferencia EstadIstica. PPU S.A., Barcelona. 21. Portenko, L.A. (1959) Studien an einigen seltenen Limicolen aus demnördlichen und östlichen Sibirien II: Der Sichelstrandläufer — Erolia ferruginea (Pontopp.). 1. Ornithol., 100, 141—172. 22. Zeiske, O. (1992) Die Rastbestände des Sichelstrandläufers Calidris ferruginea (Pont.) in den Nördlichen küstenvorländern des Elbe-Astuars — Die räumliche und zeitliche während des zuges Durch den Ostatlantischen raum. Zoologischen Institut und Zoologischen Museum der Universität Hamburg, Hamburg. 23. Stanley, P.I. & Minton, C.D.T. (1972) The unprecedented westward migration of Curlew Sandpipers in autumn 1969. Br. Birds, 65, 365—38O. 24. Hicklin, P.W. (1987) The migration of shorebirds in the Bay of Fundy. Wilson Bull., 99, 54O—57O. 25. Dunn, P.O., May, T.A. & McCollough, M.A. (1988) Length of stay and fat content of migrant Semipalmated Sandpipers in Eastern Maine. Condor, 90, 824—835. 26. Wilson, J.R., Czajkowski, M.A. & Pienkowski, M.W. (198O) The migration through Europe and wintering in West Africa of Curlew Sandpipers. Wildfowl, 31, 1O7—122. 27. Butler, R.W. & Kaiser, G.W. (1995) Migration chronology, sex ratio, and body mass of Least Sandpipers in British Columbia. Wilson Bull., 107, 413—422. 28. Iverson, G.C., Warnock, S.E., Butler, R.W., Bishop, M.A. & Warnock, N. (1996) Spring migration of Western Sandpipers along the Pacific coast of North America: a telemetry study. Condor, 98, 1O—21. 29. Butler, R.W., Kaiser, G.W. & Smith, G.E.J. (1987) Migration chronology, length of stay, sex ratio and weight of western sandpipers (Calidris mauri), on the south coast of British Columbia. 1. Field Ornithol., 58, 1O3—111. 3O. Skagen, S.K. & Knopf, F.L. (1994) Residency patterns of migrating sandpipers at a midcontinental stopover. Condor, 96, 949—958. 31. Page, G. & Middleton, A.L.A. (1972) Fat deposition in autumn migration in the Semipalmated Sandpiper. Bird-Banding, 43, 85—96. 32. Lyons, J.E. & Haig, S.M. (1995) Fat content and stopover ecology of spring migrant Semipalmated Sandpipers in South Carolina. Condor, 97, 427—437. 33. Mascher, J.W. (1966) Weight variations in resting Dunlins Calidris a. alpina on autumn migration in Sweden. Bird-Banding, 37, 1—34. 34. Szèkely, T. & Bamberger, Z. (1992) Predation of waders (Charadrii) on prey populations: an exclosure experiment. 1. Anim. Ecol., 61, 447—456. 35. Underhill, L.G., Prys-Jones, R.P., Syroechkovski, E.E., Jr., Groen, N.M., Karpov, V., Lappo, H.G., van Roomen, M.W.J., Rybkin, A., Schekkerman, H., Spiekman, H. & Summers, R.W. (1993) Breeding of waders (Charadrii) and Brent Geese Branta bernicla bernicla at Pronchishcheva Lake, northeastern Taimyr, Russia, in a peak and a decreasing lemming year. Ibis, 135, 277—292. 36. Figuerola, J. & Bertolero, A. (1996) Differential autumn migration of Curlew Sandpiper (Calidris ferruginea) through the Ebro Delta, Northeast Spain. Ardeola, 43, 169—175. 37. Pienkowski, M.W., Knight, P.J., Stanyard, D.J. & Argyle, F.B. (1976) The primary moult of waders on the Atlantic coast of Morocco. Ibis, 118, 347—365. 38. Barter, M. (1986) Sex-related differences in adult Curlew-Sandpipers Calidris ferruginea caught in Victoria. Stilt, 8, 2—8. 39. Elliot, 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 South Africa. Ostrich, 47, 191—213. 4O. Thomas, D.G. & Dartnall, A.J. (1971) Moult of the Curlew Sandpiper in relation to its annual cycle. Emu, 71, 153—158. 41. Figuerola, J. & Bertolero, A. (1995) Theoretical flight ranges of waders resting in the Ebro Delta during autumn migration. Doñana, Acta Vertebr., 22, 124—13O. 42. Lindström, A., Hasselquist, D., Bensch, S. & Grahn, M. (199O) Asymetric contest over resources for survival and migration: a field experiment with bluethroats. Anim. Behav., 40, 453—461. 43. Puttick, G.M. (1981) Sex-related differences in foraging behaviour of Curlew Sandpipers. Ornis Scand., 12, 13—17. 44. Gauthreaux, S.A., Jr. (1982) The ecology and evolution of avian migration systems. In Avian Biology 6 (eds D.S. Farner, J.R. King & K.C. Parkes), pp. 93—168. Academic Press, New York. 45. Smith, H.G. & Nilsson, J.-A. (1987) Intraspecific variation in migratory pattern of a partial migrant, the Blue Tit (Parus caeruleus): an evaluation of different hypotheses. Auk, 104, 1O9—115. 46. Puttick, G.M. (1978) The diet of the Curlew Sandpiper at Langebaan Lagoon, South Africa. Ostrich, 49, 158—167. © 1998 British Trust for Ornithology, Bird Study, 45, 313—319