Abstract. In Scandinavia Yellow Wagtails migrate with rather low average weights and small fat deposits, at most 10 - 15 % of total body weight. In Scania, S. Sweden, the mean evening weights of migrants remain stable throughout the migration period. The average distance flown per day is order of magnitude 30 - 50 km in the northern part of Sweden, 50 - 100 km in the southern third. All recoveries along the Swedish west coast, from Sweden to Denmark and from Sweden to Germany indicate a retardation when the birds meet with Kattegat, the Sound and the Baltic, the overall progress throughout this area is only 20 - 30 km/day, there is much reversed migration from Falsterbo peninsula, and possibly short scouting flights in different directions. This halt is also connected with some degree of fattening at the last major roost before Falsterbo peninsula, Foteviken, but the quality of surrounding foraging grounds may be reflected in evening weights at roosts as well. Extended stays - possibly again connected with fattening - occur at a sequence of river valleys or plains in Europe: Belgium, northern Italy and Switzerland, Iberia. The average progress within Europe is 74 km/day, this equals turning a minimum of 0. 25 g, with moderate headwinds some 0.4 g of fat/day into motion; the recorded peak performance in our material was 189 km in a single flight. The possibility that recoveries reflect the progress of a particularly slow cohort of Yellow Wagtails is discussed; under all circumstances the final flight to Africa from Italy, S. Spain and Portugal is likely to take place at a higher pace than the average pace within Europe.
In general, when there are no geographical barriers, migrants will carry small fat loads and make slow, but steady progress. Northern Bluethroats Luscinia s. svecica arrived at Kvismaren, 59°11' N almost fat-free (Lindström et al 1985), probably after a series of short flights, and fattened here before an extended flight across the Baltic and into Russia. Norman 1987 remarked that Willow Warblers Phylloscopus trochilus in autumn "migrated at weights virtually equivalent to normal roosting levels". Similar strategies have been noted in two Acrocephalus species (Bibby & Green 1981) and in inland migrants over continents (Caldwell et al. 1963, Evans 1966, Nisbet 1963).
In Yellow Wagtails wintering in Nigeria Ward 1964 recorded evening fat reserves of little more than 5 %. This enables the birds to cover overnight losses and start feeding in the morning, but not to survive the next night without extra feeding, much the same as Newton 1969 concluded for British Bullfinches Pyrrhula pyrrhula in autumn. Nigerian wintering Yellow Wagtails had mean weights 16.80 g (males) and 15.85 g (females) at dusk, average departure weights 23.9 g (males) and 22.5 g (females) and fat-free weights at water index 203 and corrected for replacement of muscle protein 15.85 g (males) and 14.85 g (females) (Wood 1982). Applying the formulae of Pennycuick 1969 Wood estimated the maximum flight distance at 2040 km, the flight lasting for some 70 hours with the birds using up on average 0.11 g of body weight per hour, a little more initially than in the end. In a 17 g bird this means a loss of 0.65 % of body weight/hour. The value 0.11 g/hour is identical with the value calculated by Nisbet et al. 1963 for Blackpoll Warblers Dendroica striata - at Bermuda lighthose fat specimens of this species weighed on average 17.3 g. Excluding the Blackpoll Warbler, Hussell & Lambert 1980 arrived at a slightly higher estimate in nine other passerine species: 0.91 % of body weight/hour.
In Belgium 147 September migrants weighed on average 17.1 g in the evening (Spaepen 1957), and Crousaz 1961 had an average evening weight of 17.0 g in late September; this seems to be the fixed mean value for North- and Mid-Europe in autumn. The fairly constant gap between morning and evening weights is emphasized by these authors and by Mester 1959 and Wood 1978. At Defilia Oasis, Morocco, Ash 1969 found average arrival weights of flava males to be 14.9 g, of flava females 13.9 g; during a cold spell in 1965 15 birds weighed on the first day averaged 13.7 g. Assuming that weights below 13.5 - 14.0 g imply exhaustion and beginning dehydration (fat-free weight at different water index levels), and assuming a 50 : 50 ratio between metabolized fat and metabolized ingested food overnight (Helms & Drury 1960), the average evening weights of wintering birds and of North European migrants indicate average fat ratios of little more than 10 %, and hence a limited radius of action in many migrants. In the present paper the above assumptions are tested on material collected in SW Scania, Sweden in autumn, and weight data are connected with a representative recovery material in order to map the general progress of migrating Yellow Wagtails on the European mainland. For material and methods see a resting draft, that will be split up on other papers, the sites and the catch are presented there.
2. Results
2.1. Weights of different categories
Tables I and II list biometry of different categories met with in SW Scania. The term "southern flava" (sfl) denotes birds from the local, two-brooded population, "northern flava" (nfl) birds with no black on ear-coverts but reduced supercilium, "flava/thunbergi" (fl/th) birds with some thunbergi characters: always black on ear-coverts but distinct supercilium. Cf. pictures of introductory paper. (For all we know the "northern flava" birds may be the really interesting category, the crucial hybrid, while "flava/thunbergi" with short supercilium behind the eye may be unadulterated thunbergi; Rendahl 1967 recorded absence of eye-stripe in only 60 % of Swedish thunbergi, and in his view the rest with short white supercilium should also be regarded as good thunbergi: "I am not convinced that Sammalistos separation is binding evidence (of hybridization)". Merikallio 1958 is quoted in support of this standpoint, contrary to Sammalisto 1958, 1961, 1968 he states that the Yellow Wagtails of Finnish Lappland are "100 % thunbergi"). Blank "1y" in the tables denotes birds attributable to a mix of all occurring subspecies.
| Category, interval | n | Mean wing ± 1 s.e.; 1 s.d. (mm) | Mean weight ± 1 s.e.; 1 s.d. (g) |
| 1y, 1-4.9 | 292 | - | 17.2 ± 0.1; 1.2 |
| ad M, 1-4.9 | 8 | - | 17.5 ± 0.4; 1.0 |
| ad F, 1-4.9 | 8 | - | 17.2 ± 0.5; 1.4 |
| 2y+ M, flava, 26.8 | 17 | 83.8 ± 0.5; 2.0 | - |
| 2y+ M, thunb., 26.8 | 6 | 83.8 ± 0.7; 1.8 | - |
| 2y+ F, 26.8 | 26 | 81.2 ± 0.4; 2.2 | - |
| 1y M, 18.8-8.9.87 | 65 | 82.7 ± 0.3; 2.1 | 17.7 ± 0.1; 1.1 |
| 1y unsexedx, 18.8-8.9.87 | 96 | 80.4 ± 0.2; 2.3 | 17.1 ± 0.1; 1.2 |
| Category, interval | n | Mean wing ± 1 s.e.; 1 s.d. (mm) | Mean weight ± 1 s.e.; 1 s.d. (g) |
| 1y M, 14.8-13.9 | 22 (31) | 83.9 ± 0.4; 2.0 | 19.4 ± 0.2; 1.3 |
| 1y F, 14.8-13.9 | 23 (28) | 80.3 ± 0.4; 2.0 | 18.1 ± 0.2; 1.1 |
| 2nd brood 1y, 13.8-14.9 | 74 (103) | 80.4 ± 0.3; 2.6 | 16.4 ± 0.1; 1.1 |
| 1y/om, 14.8-7.9 | 26 | - | 16.2 ± 0.2; 1.1 |
| 1y/sm, 13.8-1.9 | 21 | - | 16.4 ± 0.2; 0.9 |
| 1y/gm, 13.8-14.9 | 46 | - | 16.5 ± 0.2; 1.1 |
| 1y/nm, 18.8-3.9 | 10 | - | 16.9 ± 0.4; 1.1 |
| 2y+ sfl F, moulting, 6.8-13.9 | 48 | - | 16.5 ± 0.2; 1.2 |
| 2y+ sfl F, fresh, 24.8-8.9x | 23 (6) | 80.2 ± 0.5; 2.5 | 16.8 ± 0.2; 0.5 |
| 2y+ sfl M, moulting, 9.7-13.9 | 33 | - | 17.1 ± 0.1; 0.8 |
| 2y+ sfl M, fresh, 14.8-12.9x | 10 (16) | 82.9 ± 0.7; 2.1 | 17.9 ± 0.3; 1.1 |
| 2y+ flava M, 13.8-12.9 | 29 (40) | 84.8 ± 0.3; 1.7 | 18.5 ± 0.2; 1.1 |
| 2y+ fl/th M, 13.8-8.9 | 10 | 84.4 ± 0.8; 2.4 | 18.3 ± 0.3; 1.0 |
| 2y+ thunb. M, 19.8-13.9 | 25 | 83.9 ± 0.3; 1.5 | 17.9 ± 0.2; 0.8 |
| 2y+ all subsp. F, 19.8-13.9 | 20 (30) | 80.9 ± 0.4; 1.7 | 17.6 ± 0.2; 1.2 |
From the tables it is clear that Yellow Wagtail males are larger than females, and there is a correlation between weight and body-size as measured by wing length. Fig. 1 gives the regression evening weight versus wing in 164 1y birds from Löddesnäs 1987, with the regression coefficient at hand all weights can be recalculated to mean wing: 81.4 ± 0.2 mm. Mean values of the normated evening weights ± 1 s.e. in the same birds are plotted against time in Fig. 2, the overall mean value, 17.35 g, is shown with a horizontal line. The mean evening weight of juveniles is substantially higher at Foteviken, 18.56 g, and the regression is steeper; the regression evening weight versus wing in 82 1y birds (mainly from 1971 and 1987) is given in Fig. 3. The mean wing-length is also somewhat higher: 82.0 ± 0.3 mm. The regression corrected weight versus time is zero in both these cases, R-Sq(adj) = 0.0 %; average juvenile weights remain stable throughout the migration period.
Adults of both sexes and from both sites have been pooled in Fig. 4; the regression evening weight versus wing is less steep here than in juveniles from either site. With the regression coefficient at hand the weights were recalculated to mean wing: 83.6 ± 0.2 mm, normated weights are plotted versus time in Fig. 5. Again the regression is practically zero; there is no significant decline or increase in adult weights as the migration period progresses. For comparison normated evening weights of 2nd brood southern flava juveniles (mean wing 80.4 ± 0.3 mm) are plotted against time in Fig. 6; again there is no significant regression, the birds maintain their weights throughout the migration period, they are not handicapped by food shortage (and maybe in the end forced to leave the country). In Table II there is an average increase 16.2 g (old medians) - 16.4 g (shed medians) - 16.5 g (growing medians) in body-moulting juveniles, but as a matter of fact weights decreased in at least three cases where birds were caught twice: 14 Aug: old medians/17.2 g - 29 Aug: fresh medians/16.0 g; 20 Aug: old medians/17.0 g - 29 Aug: shed medians/16.3 g; 21 Aug: old medians/15.9 g - 29 Aug: shed medians/15.0 g, so there still may be some physiological stress on at least some late moulting juveniles. A moulting southern flava female on 6 August lost 1.9 g, from 15.8 to 13.9 g, in 11.5 hours, a moulting southern flava male on 14 August equally lost 1.9 g, from 17.5 to 15.6 g, in 10 hours; at least the female should have touched the upper limit of fat-free weights of females in late summer, the male limit probably lies slightly below 15.0 g. (Most of the overnight weight loss in moulting birds is due to conversion of protein; according to Newton 1968, 1972, moulting Bullfinches Pyrrhula pyrrhula lost 0.7 grams of lean dry matter each night and regained it each day; some extra weight of water must be attached to it as well. The overnight loss of metabolized fat in a wagtail weighing 17 g in August/September should be much less, we estimate: 0.3 g). Adult southern flava moulting remiges weighed much the same as 2nd brood juveniles of the same population, moulting body-feathers.
In the autumn of 2002 52 1y and 4 2y+ Yellow Wagtails were fat-ranked according to the Operation Baltic scale (e.g. Busse 2000) in Ljunghusen on the Falsterbo peninsula. The birds were caught in the beach zone late in the evening and always had left their roosts before dawn; they were migrants pushing themselves in order to leave the country. All had visible, very distinct fat deposits, still their mean weight was the same as at Löddesnäs: 17.2 ± 0.2 g, s.d. 1.3 g. Note that the steps between the four fat-classes get progressively shorter, indicating progressive dehydration as part of the preparedness for migration! The regression evening weight versus wing is shown in Fig. 7, it is steeper than the regressions recorded at Löddesnäs and Foteviken. Table III lists relevant parameters and the estimated fat content of each ranking class; no birds were assigned to fat-classes 0 and 1. A mean value (14.5 g) between the estimated female fat-free weight, c14 g, and the estimated male fat-free weight, c15 g, is used for reference, but the values of Ash 1969 indicate, that the true average level may lie closer to 14.0 g (metabolization of protein when fat is on the vane probably makes this lower limit somewhat floating). Crousaz 1961 gave a mean weight of 15.0 g for 77 "early catches" at Alp passes before September 10th and Wood 1978 had average dawn weights 14.7 - 15.0 g (sexes pooled) in Nigeria in winter. We furthermore assume that the decrease between the average evening weight and the average dawn weight is made up of 50 % metabolized fat and 50 % metabolized ingested food. Below the dawn level there is still a basic, "life-insurance" load of fat , not involved in the calculus, the "structural" lipid found by Newton 1969 in Bullfinches, it was order of magnitude 0.1 g in Bullfinches weighing 20 g and is ignored here. All estimates are minimum estimates:
| Fat class | n | Mean weight ± 1 s.e.; 1 s.d. (g) | Mean wing ± 1 s.e.; 1 s.d. (mm) | Est. fat content (g) |
| 2 | 9 | 16.1 ± 0.4; 1.2 | 81.1 ± 1.0; 2.9 | 0.8 (5 %) |
| 3 | 19 | 17.0 ± 0.3; 1.3 | 82.5 ± 0.4; 1.9 | 1.25 (7.5 %) |
| 4 | 20 | 17.5 ± 0.2; 1.1 | 82.2 ± 0.5; 2.4 | 1.5 (8.5 %) |
| 5 | 8 | 17.9 ± 0.6; 1.6 | 82.3 ± 1.2; 3.4 | 1.7 (9.5 %) |
2.3. Recoveries: distances, times elapsed
All recoveries illustrating the speed of movement in Yellow Wagtails migrating by way of S. Scania are listed in the Appendix, distances were calculated as orthodromes. Birds lagging behind in possible staging areas (Switzerland, Portugal) have been included, recoveries with vague or possibly manipulated recovery dates excluded. The three swiftest movements have been marked with red colour:
Three of our own recoveries illustrate speeds within Sweden. The material is small, but probably representative: 30 km/day from N. Sweden - possibly illustrating rather slow initial progress, and 50 - 90 km/day during peak migration from Mid-Sweden, much the same speed as on the European mainland. Two other Swedish recoveries of birds from Lappland (final list of the appendix) indicate average daily distances of the same order of magnitude: 30 and 44 km/day. Fourteen recoveries illustrate movements along the west coast of Sweden; most Yellow Wagtails leaving Sweden by way of the Falsterbo peninsula probably arrive by this "channel", many migrating flocks being deflected by the coastline until they reach the dead end of Southern Scania and are forced to cross Öresund or the Baltic. Birds from this group moved on average 176 km in 6±1 days, s.d. 3 days, mean distance flown: 30 km/day; not very impressive progress, but the obstacle, creating a slight jerk in a very regular progress, may be the Falsterbo peninsula itself. The same delay is still obvious in the recoveries from Denmark and N. Germany. A short stopover, including a few days of fattening and organization may be inserted when the birds are confronted with Öresund/the Baltic; the coastal plains of Skåne and Halland have much pasture and grazing cattle.
Between the two major catching sites, Löddesnäs and Foteviken, distance 31 km, there are 21 retraps; ringing dates 12 August - 8 September, distance 31 km, time elapsed 1 - 15 days, mean duration 5±1 days (not normally distributed); shortest durations 2x1, 7x2, 1x3 days. Most of the juveniles staying at roosts for weeks are likely to come from the local, still moulting population, e.g. Löddesnäs 20.8 - Foteviken 28.8; half of the medians still growing: southern flava. So, the mean value is not representative of the ongoing, overall migration; the shortest durations, 1 - 3 days, should be more representative. There is some degree of reversed migration as well, adding to the picture of slow progress and turmoil: three from Foteviken retrapped at Löddesnäs (30 km northbound movement, costing at least 0.11 g of fat) after 2, 6 and 9 days, 22 from Skanör and Falsterbo flying 10 km E and roosting at Foteviken, ringing dates 4.8 - 8.9, four retrapped on the day of ringing, eighteen after 1 - 13 days, mean value 5 days. Some local, moulting southern flava were involved, but others were 1y birds in complete winter plumage, and these figures fit well into the picture of delay or retardation connected with the Sound crossing.
From Denmark and North Germany there are 13 recoveries, mean distance flown 27 km/day in 6 - 7 days: one bird covered 189 km in a single flight, this is the fastest flier of the whole material, but each single flight of this kind is most probably followed by one or two days of rest, so that the average daily progress seldom exceeds 100 km. The same maximum pace can be inferred from the Belgian material; the two swiftest birds reached Belgium after flying on average 72 and 79 km/day from S. Scania. One bird reached Brescia, Italy, after flying 140 km/day for eight days, but other birds have lagged behind in Italy and Switzerland well into October. Since no recoveries follow on this course angle south of northern Italy, the birds are likely to stage here before embarking on a longer direct flight for Africa. On the SW course corresponding staging areas must be situated farther to the south, in southern Spain and southern Portugal, and possibly northern Morocco. A general feature throughout Europe is the concentration of Yellow Wagtails to river plains with pasture and cattle: Meuse and Schelde in Belgium, Ebro, Guadalquivir, Guadiana, Tejo and Douro in Iberia. The whole material from the European mainland and from Africa has been plotted as distance versus time in Fig. 8.
3. Discussion
a. Weight regressions, possible dehydration.
Starting with the regression diagrams: what causes the different regressions of Figs. 1, 3. 4 and 7? The materials obviously are somewhat irregular and of interest only as indicators of such irregularities. The fat-ranked birds from Ljunghusen in 2002 were "pushing" on the coastline in the evening; by all likelihood their migration urge was at its peak, and their fat deposits were both large and remarkably distinct. In this regression (Fig. 7) more weight increase is explained by wing-length than in any other case. The material contains 20 males and 19 females, an even sex ratio, still there is a surplus of long-winged birds, indicating some sort of selection towards such birds in the evening migration peak. (With moderate headwinds a sudden evening outbreak of migration may even evacuate the roosts at Foteviken completely; the birds probably continuing to migrate till dawn). In Table I it can be seen that 65 juveniles at Löddesnäs were provisionally sexed as males while 96 were left unsexed; sexing in the field has always indicated that there is a female surplus at this site, maybe caused by lighter and more short-winged females being less go-ahead. The whole issue depends on the reliability of sexing in juveniles; we do not know the size of our own error. If all part materials are pooled (n=446), there is an increase of 0.25 g per mm wing, this figure may be used for general corrections. The adult regression (Fig. 4) is lowest of all, and the prediction of increasing weight by wing-length least; here there is reason to suspect some degree of dehydration (cf Wood 1982) combined with rather low fat deposits: a pronounced lightweight strategy. We believe this capacity to be best developed in older birds.
b. Stable mean weights: throughout the migration period in Scania, over latitudes in Europe.
The mean weights of Spaepen 1957 and Crousaz 1961 were confirmed at Löddesnäs (1-4.9) and at Ljunghusen in 2002; we conclude that the evening mean weights are stable over at least 9 degrees = 1000 km in Europe. On the other hand all categories roosting at Foteviken except the local population (adults and 1y alike) will at times weigh 1 - 1.5 g more. Fig. 2 illustrates the amplitude in a single year; variations in insect supply may be the cause, drought or wind the ultimate factors. Some degree of forced staging may be involved as well; the regular flow of migration over Falsterbo peninsula is easily quenched by high temperatures, unfavourable wind directions, poor visibility; the number of roosting birds at Foteviken may exceed 10.000 during such periods. Still, in all categories and at all sites weights have remained stable throughout the migration period. In different warblers studied at Ottenby Åkesson et al. 1995 had both species with increasing and species with decreasing weights/fat rankings as the migration period progressed. Some 350 km to the south, at Berlin, Germany, Dittberner & Dittberner 1984 had increasing weights in late September and October (small samples), however. Such fattening would be adequate to the flocks staging in Italy, Spain and Portugal by the same time, it may simply be governed by an internal timer; in Scandinavia numbers are dwindling at that time and the prospect of substantial fattening here is of no practical importance.
c. Fat-rankings indicate dehydration as well.
Next: why are the distances between the fat-classes 2 - 5 (0.9 g - 0.5 g - 0.4 g) of Table III not equal? In a well-conceived ranking scale the distances between fat-classes should be much the same, and if this is not the case, the scale should be modified. We have found the Operation Baltic scale to give coherent results with e.g. Reed Buntings Emberiza schoeniclus and Blue Tits Parus caeruleus, but in the case of Yellow Wagtails in autumn, birds with large fat deposits do not weigh as much as they "should". This question can not be settled by this observation alone, but the existing measurements indicate, that some degree of extra dehydration is connected with the higher fat-classes in the Yellow Wagtail. Wood 1982 discussed this option at some length; in spring birds a reduction of the water index from 203 to 150 will increase the radius of action by some 450 km, or 20 hours of flight. It should be added, that fat-ranking may be highly subjective and irreliable: members of a team should constantly calibrate their methods against each other. In the future we intend to compare results of the slightly deviating Ottenby scale (Pettersson 1983) with results of the Operation Baltic scale, to see if the discrepancies are scale-related.
d. Overnight weight decrease: migration and roost weight losses much the same.
When it comes to overnight weight losses, Stolt & Mascher 1962 assumed the ratio 50:50 between metabolized fat and ingested food suggested by Helms & Drury 1960 to be valid in Bluethroats Luscinia svecica as well. In winter Bullfinches Pyrrhula pyrrhula covered only 20 - 25 % of their energy requirements by metabolizing seeds stored in the gullet, however, the rest was due to metabolized fat (Newton 1969). The Yellow Wagtail is exclusively insectivorous; the assumed ratio would oblige a bird to carry 0.5 - 1.5 g of ingested insect food at dusk. We expect the overnight weight loss of roosting migrants over nine hours (21h - 06h) to be at least 50 % less than the recorded overnight loss of moulting adults (1.9 g); in the latter case we estimate the overnight loss of protein + water at order of magnitude 1 gram (Newton 1968). The other half is due to metabolized fat, metabolized ingested food and excretions. At Swiss Alp passes Crousaz 1961 had surprisingly low average "early" weights: 15.0 g before 10 September (n=77), increasing to 15.9 g 11 - 25 September (n=62) and 16.1 g after 25 September (n=26); in contrast average evening weights were normal 17.0 g 11 - 25 September (n=52) and after 25 September (n=46). By all likelihood the late September birds had performed some migration during dark hours, using up order of magnitude 1 g of body weight. Like the Willow Warbler Phylloscopus trochilus (Norman 1987). Yellow Wagtails migrate at normal roosting weight levels in north and central Europe, and morning weights do not differ much here, whether the bird has flown some distance or simply roosted; the migration flight comes to a halt at some critical level before the fat deposit is empty.
e. All Yellow Wagtails should be able to manage the average pace.
From Table III it is clear that females weighing 18 g and males weighing 19 g must have fat deposits approaching 15 % of total body weight. Still, there is little in recoveries to indicate a higher migration pace following S. Scania and possibly connected with such birds; seven birds were retrapped at Gedser, 116 km SSW, on average one week later, another seven took on average 6 days to reach roosts in North Germany, and the fastest flier could have flown its 189 kms in 6 hours, spending a mere 0.7 g of fat if it departed with an initial weight of 18 g. The same slow rate of progress is obvious already along the Swedish west coast, fourteen birds using on average 6 - 7 days to cover 200 km, with 90 km in a single day as peak performance. Even birds weighing 16 g - and there are many of them, from replacement or true 2nd broods - should be able to match this rate of progress; they need not do without the protection against predation offered by large flocks. Yellow Wagtails don't overexert themselves when flying from breeding-grounds to 50° N in Europe; migrating birds are at no stage at the end of their resources and the weight composition of flocks is not uniform. It seems to us as if the lightweight individuals are setting the pace of the overall progress.
f. The final flight to Africa from staging areas in South Europe.
Two features strike the eye in Fig. 8: the very regular, linear overall progress and the accumulation (and temporal "stretching") of recoveries at certain levels: Belgium, Italy, N. Spain, S. Spain + Portugal. On one hand it is quite possible that belated and poorly developed juveniles may fly 75 km/day all the way to W. Africa, never exceeding this limit, on the other hand - challenging the tempting linearity of Fig. 8 - it cannot be excluded that some Yellow Wagtails, maybe in particular experienced adults, rest for some time and then fly nonstop to the next major staging area, in this way creating a more zigzag progress distance versus time, with faster in-between transports. Dittberner & Dittberner 1984 mention a bird from Rüdersdorf, Berlin on September 3rd to North Italy 5 September; some 750 kms flown within two days if the finding details are correct. The zigzag hypothesis may in particular hold true of the SW course by way of France and Iberia; in the end there may even follow a substantial fattening and a loop over the Atlantic as asserted by Gatter 1987 (in translation, references omitted):
Many Palaearctic migrants do not leave the SW corner of Portugal (e.g. Cap Vicente) on a southerly direction, in order to reach the Moroccan coast as quick as possible. Instead they maintain the SSW course till well over the open sea. Regular passage of Palaearctic migrants on the Canary Isles is a wellknown fact (a population of Eleonora Falcons, Falco eleonorae, depends on it for its breeding). On the other hand resting Palaearctic migrants are not numerous in the isles. Therefore it can be expected, that many migrants reach the coastline of Africa no sooner than at Cap Blanco or even farther to the south, and cross the Sahara over land only in its southern part.
The bird reaching Nouakchott after 27 days (on average 177 km/day for almost a month) should serve as a memento in an overwhelming context of apparent slow, linear progress; if the recovery details are correct, it most probably flew to Portugal in a fortnight, fattened here and then flew nonstop to Mauritania. The absence of recoveries on the SSW course after N. Italy is a problem; Moreau 1972 suggested that winds from northerly directions are likely to assist migrants on this route in autumn, so the fat deposits necessary for a safe Sahara crossing will be less than in spring.
g. Is our material representative?
It is clear from the above that biometry, physiological state and maybe hierarchic position in flocks, dominance, are likely to affect the behaviour of Yellow Wagtails, their choice of roost-sites, their flight performance, their need of refueling. Which birds are primarily sampled when we ring at roosts? Could it be that we at one site (e.g. Löddesnäs) sample primarily from birds in need of refueling, birds restraining themselves for some reason, while at another site (e.g. Ljunghusen) the daring, go-ahead birds will be in majority and most caught? And which category is most likely to be the victim of exhaustion or to be caught by other ringers? (Of course separation, if there is some, will never be complete, but only partial). This question is primarily brought up by the flight from Scania to Mauritania in less than a month, but also by a few other recoveries on the left hand of Fig. 8 and by the recovery void between northern Italy and Africa. Is the linear, 75-km-per-day progress presented here as a general feature of Yellow Wagtails, when it comes to it only typical of a limited cohort of birds, while another cohort migrates faster and hence has less probability of recovery? We do not have the answer, but we think, that the question should be raised.
Summary:
1. Mean evening weights of migrating 1y Yellow Wagtails in SW Scania were 17.2±0.1 g at both Löddesnäs 1 - 4 September and Ljunghusen 28 August - 10 September. On the other hand 31 1y males from the major roost Foteviken 14 August - 13 September weighed 19.4±0.2 g, while 28 1y females weighed 18.1 ± 0.2 g. Fattening could be a reaction to topological obstacles (the Sound, the Baltic), at the same time it could reflect the quality of foraging-grounds and forced resting due to unfavourable weather: too warm, too windy, poor visibility (cf. high and low weights of Fig. 2). Mean weights remained the same throughout the migration period, as a rule the birds do not experience shortage of insect food (mainly Diptera species) in August and the first half of September.
2. There were no fat-free birds among 56 Yellow Wagtails ranked in 2002, the four fat-classes recorded should represent mean fat deposits of at least some 6.5, 9, 10 and 11 % of body-weight. The fatter birds may be more dehydrated, however, and the mean weight of fat-class 5, 17.9 g, should imply as much as 15 % fat in dehydrated birds.
3. The longest non-stop migration illustrated by recoveries was 189 km in one day, from Scania to Germany. This flight could have been achieved in 6 hours, using up 0.7 g of fat. Judged from recoveries average migrating Yellow Wagtails are not even close to depleting their fat reserves within Europe; even birds weighing only 16 g should be able to keep up with the overall pace of movement. This is an important observation: even retarded juveniles need not do without the protection against predation offered by large congregations of birds.
4. The average progress of Yellow Wagtails within Sweden is some 30 km/day in the northern part of the country with 50 - 90 km/day as peak performance in the southern part. Along the Swedish west coast and between Scania and Denmark/Germany there is a clear retardation: the average progress again amounting to some 30 km/day. The best performing migrants flew 140 - 190 km/day over extended periods after entering the European mainland.
5. On the European mainland the pattern of recoveries shows a distinct linear progress with on average 74 kms flown per day. There are at least four major staging areas: Belgium, Northern Italy, Northern Spain and Southern Spain + Portugal. The final flight to Africa from Italy and Iberia probably involves higher fat levels and longer non-stop flights, but it should be aided by tailwinds and less arduous than the Sahara crossings in spring.
APPENDIX:
From N. and central Sweden (n=3):