JOURNAL OF AVIAN BIOLOGY 33: 15–22, 2002

Ecological barriers shaping fuel stores in barn swallows Hirundo rustica following the central and western Mediterranean flyways Diego Rubolini, Andrea Gardiazabal Pastor, Andrea Pilastro and Fernando Spina

Rubolini, D., Gardiazabal Pastor, A., Pilastro, A. and Spina, F. 2002. Ecological barriers shaping fuel stores in barn swallows Hirundo rustica following the central and western Mediterranean flyways. – J. Avian Biol. 33: 15 – 22. The crossing of ecological barriers is among the most energy-demanding and risky phases of migration. Scanty field data exist on the relationships between ecological barriers and pre-migratory fuel storing in songbirds. The aim of this study was to analyse whether the distance to be covered across ecological barriers can be considered as a factor affecting pre-migratory fuel stores in barn swallows Hirundo rustica following the two main western European flyways, funnelling through Iberia and Italy. Data refer to 13029 swallows and were collected during July– October in 1997–1998 at 19 roost sites, scattered over Spain and Italy (south of 43°N). During the post-breeding phase (PB, July– August) energy stores did not differ significantly between the two geographical areas, whereas during the pre-migratory phase (PM, September– October) swallows in Italy carried larger energy stores than those in Spain. Assuming that swallows leave for migration from the fuelling site, we found a significant positive correlation between the width of ecological barriers and an average index of energy stores for each roost site during the PM phase. The width of ecological barriers (along a N-S migration route) was expressed as (1) the distance between the roost site and the coast of North Africa (representing the Mediterranean Sea), and (2) the distance between the roost site and the southern margin of the Sahara desert (the total width of ecological barriers). The weaker correlations obtained when considering only the Mediterranean as a barrier suggest that swallows may cross the desert without substantial refuelling in North Africa. Hence, we showed that fuel stores have a degree of population-specific variability among Italian and Iberian barn swallows and that the extension of ecological barriers may play a role in determining the amount of stores needed for the migratory flight. D. Rubolini and F. Spina (correspondence), Istituto Nazionale per la Fauna Sel6atica, 6ia Ca’ Fornacetta 9, I-40064 Ozzano Emilia (BO), Italy. E-mail: infsmigr@ iperbole.bologna.it. A. Gardiazabal Pastor, BIOMA T.B.C., C/Ceuta 10, E-28292 Galapagar/Madrid, Spain. A. Pilastro, Dipartimento di Biologia, Uni6ersita` di Pado6a, 6ia U. Bassi 58 /B, I-35131 Pado6a, Italy.

Western European long-distance passerine migrants have to cross both the Mediterranean Sea and the Sahara desert during their autumn migration to Africa. Most species (in particular sylviid warblers and flycatchers) leave the breeding grounds in late summer with low or modest fat depots, gradually increasing their energy stores and body mass along the migratory route until they are theoretically able to embark on a sustained flight across these ecological barriers (Moreau and Dolp 1970, Bibby and Green 1981, Veiga 1986, Bairlein 1991, Kaiser 1992, Berthold 1993, Schaub and Jenni 2000a, b). Whether passerine migrants cross the

©

barriers in a single sustained flight or in intermittent flight bouts and daytime stopovers in the desert remains the subject of much speculation, although in many species there is evidence for mixed strategies (Moreau 1972, Biebach 1992). The proportion of individuals adopting a given strategy may depend on a complex interplay of endogenous and exogenous factors: likely, migratory birds show a flexible response to their internal state and to environmental conditions (Biebach 1990, Biebach et al. 2000). Due to its aerial feeding habits, not tied to particular stopover habitats as other long distance migrants, the

JOURNAL OF AVIAN BIOLOGY

JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

15

barn swallow Hirundo rustica has been considered to adopt a fly-and-forage migration strategy, refuelling en route and thus being able to cross ecological barriers with a low or moderate amount of energy stores (Davis 1965, Mead 1983, Ormerod 1989, 1991, Cramp 1998). However, recent studies have shown that the amount of pre-migratory fattening in the barn swallow is comparable to that of migrants that adopt a strategy of direct crossing of ecological barriers in long-distance flights (Berthold 1975). Fuelling in this species commonly reaches 30 –40% of lean body mass, which, according to recent calculations, may allow uninterrupted flights of over 3000 km (Pilastro and Magnani 1997, Pilastro and Spina 1999). Hence, non-stop flights across ecological barriers may be the rule in this species (e.g. Moreau 1961), considering also that desert stopover of swallows has been reported only rarely (e.g. Biebach 1990, Biebach et al. 1991, Bairlein 1992, but see Bairlein 1985, 1988). In recent years, a growing interest in barn swallow migration in Europe, stimulated by the EURING Swallow Project, has led to the description of spatio-temporal patterns of autumn pre-migratory fattening, using a large data-set collected across 18 different countries in Europe (Jenni 1998, Spina 1998). Preliminary analyses have shown that conspicuous fuelling takes place from the beginning of September; fatter birds are found mostly in southern Europe, in particular Spain and central and southern Italy (south of 43°N) (Pilastro and Spina 1999, Rubolini 2000). Population-specific heritable variability in migratory behaviour has been demonstrated in captive, handreared passerine migrants (e.g. blackcap Syl6ia atricapilla, Berthold et al. 1992, Berthold 1996). However, there is a lack of field data demonstrating the existence of population-specific pre-migratory fuelling dynamics that could be related to the distance to be covered between breeding/fuelling and wintering areas (e.g. Bairlein 1991, Drent and Piersma 1990). Here we quantified fuel stores (sensu King and Murphy 1985, Lindstro¨ m and Piersma 1993) in barn swallows before autumn migration at the northern edge of the ecological barriers represented by the Mediterranean Sea and the Sahara desert. The aim of this study was to analyse whether ecological barriers can be considered among the selective factors affecting the amount of pre-migratory fuel stores in the barn swallow along the two main western European flyways, namely those funnelling along Iberia and Italy. In fact, swallows seem to cross the Mediterranean Sea and the Sahara desert rather than to circumfly these barriers: if so, and if the width of these two barriers varies, we expected energy stores at departure for migration to be directly related to the width of the barriers. Hence, (1) fuel stores during the pre-migratory phase (September –October) among swallows studied in central and southern Italy should be larger than among Iberian birds, given the wider eco16

logical barriers to be crossed by the former. If birds are leaving the roost areas directly for the trans-Saharan flight, and assuming a N-S migratory route, we expected (2) fuel stores at departure to be positively related to the linear distance to be covered.

Materials and methods Data selection, variables and data analysis Data used for this study were collected during the period July –October in 1997 and 1998 at a total of 19 roost sites, 8 of which are located in Spain and 11 in central and southern Italy (C&S Italy hereafter, defined as south of 43°N – i.e. the northernmost latitude of the Spanish roost sites – following the definition given in Pilastro and Spina 1999) (Table 1, Fig. 1). The selection of Italian roost sites south of 43°N allowed us to compare fuelling dynamics in terms of latitudinal range: mean latitude for Italian and Spanish roost sites did not differ significantly (Mann-Whitney U test, U = 22, P = 0.14). Swallows were tape-lured and captured with mistnets at dusk, starting from 3 h before sunset; birds were ringed and measured following standardised field protocols (Jenni 1998) and were usually released the morning following capture. For most individuals data recorded included: hour of measurement (local solar time), age (juvenile, i.e. born in the same calendar year, or adult) (Svensson 1992), length of the 8th outermost primary (hereafter wing length, nearest 0.5 mm) (Berthold and Friedrich 1979, Jenni and Winkler 1989) fat score (on a 0 –8 scale, Kaiser 1993), moult score (on a 0–2 scale, with 0 =no growing body feathers, 1 = 1– 20 growing body feathers and 2=\20 growing feathers) and body mass (at least to the nearest 0.5 g). The hour of measurement was standardised with respect to time of sunset for every roost site and trapping session, using the following transformation: standardised hour of measurement (STM hereafter) = (hour of measurement −time of sunset). Time of sunset was calculated using the Mpj Astro 1.5.1 software. Birds in heavy moult of body feathers ( \20 growing feathers) and with growing or moulted wing feathers were excluded from analyses (given that body feather moult may negatively affect fat accumulation and body composition, Newton 1968, Chilgren 1977, Spina and Massi 1992, Murphy 1996). Fat scoring provides a reliable index of the lipid stores of passerine birds (Redfern et al. 2000) and is useful for intra- and interspecific comparison of fat stores, since it is independent of size, stomach, gut and water content of the body (Brown 1996). In the analyses fat score was ln(1 + fat score)transformed, in order to improve normality (Brown 1996). The autumn pre-migratory fattening period of the barn swallow was divided into two distinct phases JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

(Pilastro and Magnani 1997, Pilastro and Spina 1999): (1) a post-breeding phase (PB hereafter) (July to the end of August), during which birds terminate the partial body moult and energy stores are low; and (2) a pre-migratory phase (PM hereafter) (from the beginning of September onwards), during which a substantial fuelling takes place (Fig. 2). We defined the PB phase from 1 July to 19 August and the PM one from 30 August to 31 October; data collected between 20 and 29 August, a 10-day period of transition between the two phases, were excluded from the analyses. Recaptures were not considered. Sample size may vary between different analyses because of missing data from some sites. Overall, the data set consisted of 13029 birds, 3730 of which belong to the PB phase (7 sites) and 9299 to the PM phase (18 sites) (Table 1). Statistical analyses were carried out using ANCOVA models and Spearman Rank Correlation test, using SPSS ver. 10.0. Exploratory analyses of factors affecting barn swallow fuel stores in southern Europe have shown them to be linearly related to both the latitude and longitude of the roost site (Pilastro and Spina 1999). These variables were therefore used as covariates in the following ANCOVA models, to represent an effect of roost site: however, in the analyses where the geographical areas (C&S Italy and Spain, representing a longitude component) were the factors, we omitted longitude from the covariates, to avoid redundancy of confounding variables. Interaction terms were included in models and presented when relevant.

Comparison of energy stores between C&S Italy and Spain Fat score and body mass were compared between the two geographical areas, separately for adult and juvenile birds and for the PB and PM phases. In the ANCOVA, body mass or fat score were the dependent variables, year and geographical area (C&S Italy and Spain) factors, while date of capture (day 1 = 1 July), date2, STM, latitude of roost site (in centesimal degrees), and wing length were entered as covariates. Date2 was included in the model to control for non-linear patterns of fuelling. Wing length was considered as a measure of body size. The inclusion of wing length in the models was needed in order to control for both individual and geographical body size differences. Mean wing length (calculated for at least 10 individuals) for a given site was positively related to the longitude of the site (juveniles, rs = 0.88, n = 17, P B0.0001; adults, rs = 0.77, n = 16, P B0.0001), swallows ringed in Italy being on average 2 mm larger than those recorded in Spain.

Fat stores of adults and juveniles Differences in fat stores between adult and juveniles were tested separately for the two geographical areas. In the ANCOVAs, fat score was the dependent variable, while age (adult and juvenile), fattening phase (PB and PM) and year were factors, and date of capture,

Table 1. Geographical location of roost sites included in the analysis and number of swallows ringed at each site. Italian roost sites were located in central and southern Italy (south of 43°N, Pilastro and Spina 1999). PB phase = post-breeding phase (1 July–19 August); PM phase = pre-migratory phase (30 August–31 October). The width of barriers is the distance (in km) between the roost site and the coast of North Africa (sea) or between the roost site and the southern margin of the Sahara desert (overall), according to a vegetation map (TCI 1998). Site

1. Madonna del Porto 2. Foce del Piomba 3. Aiguamolls de l’Emporda` 4. Lesina 5. Macchiagrande 6. Serre Persano 7. Le Cesine 8. Alimini Piccolo 9. Laguna de San Juan 10. Castillejo 11. Borriana 12. Legnochimica Rende 13. Molentargius 14. Palermo 15. Rio Argos 16. Gorghi Tondi 17. La Margazuela 18. Campillos 19. Las Cabezas de San Juan

Country

Coordinates

Operating years

Italy Italy Spain

42°35%N 12°15%E 42°32%N 14°10%E 42°00%N 03°00%E

98 98 97

Italy Italy Italy Italy Italy Spain Spain Spain Italy Italy Italy Spain Italy Spain Spain Spain

41°54%N 41°52%N 40°34%N 40°21%N 40°11%N 40°08%N 39°57%N 39°54%N 39°22%N 39°14%N 38°07%N 38°06%N 37°37%N 37°17%N 37°02%N 36°59%N

97 97–98 98 97 97–98 97–98 97–98 98 97–98 97–98 98 98 98 97 98 98

JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

15°34%E 12°17%E 15°06%E 18°20%E 18°27%E 03°31%W 03°44%W 00°05%W 16°13%E 09°09%E 13°22%E 01°48%W 12°40%E 06°02%W 02°10%W 05°56%W

Number of swallows

Width of barriers

PB phase

PM phase

sea

overall

230 – 170

278 1299 158

990 1053 599

2988 2970 2934

– 769 – – – 1760 766 – – 24 – – – – 11 –

566 852 692 418 392 1397 811 161 296 686 276 24 86 387 – 520

1098 936 855 1017 1008 551 536 466 819 232 549 333 522 191 – 131

2898 2916 2736 2808 2808 2466 2466 2610 2628 2520 2520 2232 2466 2196 – 2214

17

Fig. 2. Mean daily mass (calculated for each site for days with at least 5 individuals) in (a) adult and (b) juvenile barn wallows. PB =post-breeding phase (1 July – 19 August), PM = pre-migratory phase (30 August – 31 October). The shaded areas indicate a 10-day period of transition between the two phases. =C&S Italy,  =Spain. Fig. 1. (a) Location of ringing sites (see Table 1 for site description); (b) width of the Mediterranean Sea and Sahara desert according to a vegetation map (TCI 1998); solid vertical lines are examples of the width of ecological barriers; line (I) shows hypothetical migration distances between an Italian roost site and the coast of North Africa (width of sea), and line (II) between a Spanish roost site and the southern border of the Sahara desert (width of total barriers). Hypothetical migration distances were calculated assuming a N-S migratory route (see Materials and methods for details).

date2, STM, latitude and longitude of roost site (in centesimal degrees) were covariates. In order to further look at geographical and age differences in fuel loads, we estimated lean body mass (LBM) as the average

mass of birds with fat = 0 (Pilastro and Spina 1999). Since LBM did not differ between Italy and Spain or between age classes (two-way ANOVA, F3,425 = 0.95, P= 0.42), a single mean value of 17.9 g was used. To obtain a rough estimate of the amount of stores carried by individual birds ready for departure, we subtracted LBM from the body mass of the 25% heaviest birds (e.g. Biebach 1990, Bairlein 1991) (selected separately for each geographical area and age class) during the PM phase. We defined this value as the estimated departure fuel load (EDFL thereafter, expressed as a % of LBM) (Table 2). EDFL was entered into an ANCOVA in which year, geographical area (C&S Italy and

Table 2. Mean values of body mass, fat score and EDFL (Estimated Departure Fuel Load, see Materials and methods for details) for barn swallows during the pre-migratory phase (30 August to 31 October); values in parentheses are SD and sample sizes. Geographical area C&S Italy Adults Juveniles Spain Adults Juveniles

18

Body mass

Fat score

EDFL

22.5 (2.5; 858) 21.1 (2.3; 4925)

4.1 (1.1; 869) 3.9 (1.1; 4927)

43.8 (6.2; 220) 35.7 (5.8; 1233)

21.1 (2.3; 320) 19.6 (2.1; 3101)

3.2 (1.4; 322) 2.9 (1.3; 3099)

33.9 (5.7; 80) 24.6 (6.5; 841)

JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

Results Adult and juvenile fuel stores in Spain and Italy

Fig. 3. Mean fat score of adult and juvenile barn swallows during the PB and PM phases in (a) C&S Italy and (b) Spain.

Spain) and age were factors, and date, date2, STM and latitude were covariates.

Energy stores and width of ecological barriers For every roost site in Italy and Spain during the PM phase we obtained an index of energy stores, separately for adult and young birds. The index was derived from the residuals of an ANCOVA in which fat and body mass were entered as dependent variables, year as factor, and date of capture, date2, STM and wing length were considered as covariates. The average of residuals for each roost site (for at least 10 individuals) was considered as an index of the energy stores at departure, assuming that swallows leave for migration from the fuelling site. This index was correlated with the width of ecological barriers. To calculate the latter, we measured the linear distance (in km) between the roost site and the coast of North Africa (representing the barrier formed by the Mediterranean Sea), and the distance between the roost site and the southern margin of the Sahara desert (representing the overall width of ecological barriers) as indicated on a vegetation map (TCI 1998). Distances were measured assuming a N-S migration route. A sequential Bonferrroni correction (Rice 1989) was used to correct P-values for multiple testing (on a total of 8 correlations). JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

During the PB phase, energy stores did not differ significantly between the two geographical areas, either in terms of fat score or body mass (fat score: adults F1,216 = 3.23, P =0.07; juveniles F1,3417 = 0.67, P =0.41; body mass: adults F1,216 = 0.15, P = 0.70; juveniles F1,3410 = 0.46, P =0.49). During the PM phase birds were fatter in Italy (fat score: adults F1,1107 = 16.5, PB 0.0001; juveniles F1,7392 = 184.6, P B 0.0001; body mass: adults F1,1097 = 6.3, P = 0.012; juveniles F1,7392 = 75.4, P B 0.0001) (Table 2). Young birds were fatter than adults during the PB phase, whereas adults had significantly larger fat stores than juveniles during the PM phase: the same pattern was recorded in both geographical areas (C&S Italy, age ×phase interaction, F1,6805 = 15.9, P B 0.0001; Spain, age ×phase interaction, F1,6057 = 30.8, P B 0.0001) (Fig. 3). Although differences in mean fat score were significant, the differences in the absolute values during the PM phase were very small (Table 2): however, fat score is an index of the amount of fat stores, and by no means can be interpreted as the true amount of fat carried by an individual. Nevertheless, when considering EDFL, adults departed with significantly greater fuel loads than juveniles (41.2% vs 31.2% LBM, F1,2373 = 455, PB 0.0001), and Italian-ringed birds with significantly greater fuel loads than Spanish-ringed ones (36.9% vs. 25.4% LBM, F1,2373 = 393, P B 0.0001). Overall, the difference in fuel loads between Italy and Spain was 11.1% among juveniles and 9.9% among adults (Table 2).

Energy stores and ecological barriers The ANCOVAs from which we obtained residuals of energy stores were highly significant for both age classes (fat score: adults F5,1116 = 30.9, R2 = 0.11, P B 0.0001; juveniles F5,7401 = 236.1, R2 = 0.13, P B 0.0001; body mass: adults F5,1106 = 69.9, R2 = 0.24, P B 0.0001; juveniles F5,7392 = 552.4, R2 = 0.27, P B 0.0001). For both body mass and fat score the condition index at every trapping site was independent of the number of trapped swallows (Spearman Rank, adults: P \0.1; juveniles: P \ 0.4). The correlations between the two condition indices and the width of ecological barriers were positive and significant (Table 3, Fig. 4). The residuals for juvenile and adult birds were highly correlated (body mass: rs = 0.83, n = 16, P B 0.0001; fat: rs = 0.92, n = 16, P B0.0001). The correlations between fuel stores at departure and width of ecological barriers might have been confounded by pooling data from Iberian and Italian birds. In fact, correlations may arise because of different ecological conditions affecting 19

fuelling dynamics in the two geographical areas, hence causing a spurious relationship. The positive correlation between fuel stores and the overall width of the barriers still holds – at least for body mass data – even ‘‘within’’ each geographical area, despite the smaller sample size and reduced variability in barrier width, thus further supporting our results (considering the average value of adult and juvenile residuals from ANCOVAs for each roost site, body mass: C&S Italy rs =0.63, n =10, P = 0.05; Spain rs =0.85, n =7, P= 0.02; fat score: C&S Italy rs =0.59, n = 10, P = 0.07; Spain rs =0.59, n =7, P=0.15).

Discussion Previous analyses of barn swallow body mass variation across Europe have indicated that important fuelling areas are located around the Mediterranean, particularly in Italy and Iberia (Pilastro and Spina 1999, Rubolini 2000). These areas host the last stopover sites before the journey across the ecological barriers represented by the Mediterranean Sea and the Sahara desert. Western European populations leave the breeding grounds in continental and northern Europe with low energy stores, complete their post-breeding and post-juvenile body moult during the earlier part of the trip, reach the main fattening areas of southern Europe during August and then, after completion of the body moult, start rapidly to gain weight and depart for the African wintering grounds in September (Pilastro and Magnani 1997, Pilastro and Spina 1999).

Table 3. Correlation (Spearman rank correlation coefficient, rs) between the mean index of energy stores at each roost site (calculated for at least 10 birds during the PM phase, see Materials and methods for details of index calculations) and: (a) the width of the Mediterranean Sea, and (b) the total width of ecological barriers (Mediterranean Sea and Sahara desert). Sample size (n) refers to the number of roost sites. Migration distances were measured assuming a N-S migratory route. All P-values significant at PB0.05 after sequential Bonferroni correction (Rice 1989).

(a) Mediterranean Sea Adults Body mass Fat score Juveniles Body mass Fat score (b) Total barriers Adults Body mass Fat score Juveniles Body mass Fat score

20

rs

n

P

0.61 0.55

16 16

0.012 0.027

0.72 0.59

17 17

0.001 0.013

0.64 0.68

16 16

0.007 0.004

0.75 0.72

17 17

0.001 0.001

Fig. 4. Mean residuals from ANCOVA models for (a) body mass and (b) fat score for each trapping site (with sample size ]10 individuals) in relation to the linear distance between the trapping site and the southern margin of the Sahara desert (see Materials and methods for further details). Triangles = adults, circles =juveniles; filled symbols = C&S Italian sites, open symbols =Iberian sites; linear regressions are shown (broken line =adults; solid line =juveniles).

The indices of energy stores for each of our study sites were significantly correlated between the two age classes, indicating that adults and juveniles build up energy stores in a similar way at each site, although at the end of the fuelling period (during the PM phase) juveniles attain a lower level of energy stores both in Spain and Italy. A reverse difference in physical condition between age classes is recorded during the PB phase, when adults show a more intense body feather moult (Pilastro and Magnani 1997, D. Rubolini, A. Massi and F. Spina, unpubl. data): juveniles were found to have relatively larger stores. The larger energy stores in adults at departure may arise from a faster and more efficient fuel deposition rate among adults as compared to juveniles and/or from a higher foraging efficiency, given that fuel accumulation begins almost concurrently in the two age groups and geographical areas (see Fig. 2) (Pilastro and Magnani 1997). The greater fat loads among adults is likely to increase their JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

chances for a successful termination of the migratory journey, because swallows seem to experience a relatively high mortality during migration compared to other species of trans-Saharan migrants (Moreau 1961). While no geographical differences in the level of fuel stores where found during the PB phase, our results show that during the PM phase barn swallows fattening in Spain accumulated less fuel than those fattening in central and southern Italy. When considering EDFL and mean body mass, such a difference is remarkable: both adult and juvenile swallows leaving from Italy carried about 1.5 –2 g of fuel (10% of LBM) more than those from Spain. Recently, Pilastro and Spina (1999) estimated potential flight ranges for barn swallows, following Pennycuick (1996). According to their results, a 2 g difference in fuel stores means that the birds departing from Italy can potentially fly about 800 km longer than those leaving from Spain, although the difference in the overall width of ecological barriers is only 400 km. A possible explanation for this apparent over-fuelling recorded in Italian birds may lie in the different width of the stretch of sea to be crossed (800 vs 400 km of sea on average for Italian and Spanish birds, respectively). Sea crossing may impose higher energetic costs on migrating swallows, because weather conditions over the sea may be more unpredictable than over the desert, where northerly tail-winds that facilitate the migratory flight prevail during autumn (Moreau 1961, Biebach 1992). In fact, Moreau (1961) reported that, although winds over the Mediterranean are more favourable in autumn than in spring, wind direction and intensity over the sea are indeed more variable than over the desert. As predicted, the amount of fuel stored at each roost site was significantly correlated with the extension of the ecological barriers to be crossed, under the assumption of a due south migratory direction, i.e. swallows leaving for their trans-Saharan flight along the shortest route. This relationship was weaker when considering only the Mediterranean as a barrier, while it was stronger when taking into account both the Mediterranean Sea and the Sahara desert, suggesting that swallows probably cross the desert without substantial refuelling in North Africa. Whether this strategy is adopted by all individuals or only by some of them remains an open question. Further, we cannot exclude the possibility that at least a portion of Iberian swallows may follow the West African coast, while direct barrier crossing by Italian birds is suggested by a single recovery of a barn swallow reported 7 days after ringing from Niger, having covered a total of 3028 km (433 km/day) (Pilastro and Spina 1999). However, along the Atlantic coast of Morocco, southward migrating swallows were found in very low numbers compared to the interior of that country (Moreau 1961), suggesting that the number of Iberian swallows following the border of the desert is probably low. JOURNAL OF AVIAN BIOLOGY 33:1 (2002)

The differences between Italy and Iberia in the amount of fuel stored by swallows is comparable to that recorded in another trans-Saharan migrant, the garden warbler Syl6ia borin during autumn migration (Bairlein 1991). In this species, body mass values are higher in southern Europe (just prior to barrier crossing) than in northern Europe, and in southern Europe birds migrating along the eastern flyway are about 3 g heavier than those following the western flyway (Bairlein 1991). Interestingly, the highest body mass values among southward migrating garden warblers were recorded at stopover sites in northern Sahara (Bairlein 1991), while body masses of swallows roosting in Algeria were similar to those recorded in southern Europe (Bairlein 1988), again suggesting that no substantial fuel stores are accumulated in North Africa. The lower amount of fuel stored by the Iberian-fuelling population seems to be in line with the theoretical ecological and fitness costs of carrying extra fuel loads that may, e.g., increase predation risk by impairing flying ability, or cause extra energetic costs for flying (Lima 1986, Witter and Cuthill 1993, McNamara et al. 1994). Thus, the observed differences in fuel stores between the two populations may indicate that selection has led to the optimisation of fuel loads at the population level, hence providing indirect evidence that carrying fat imposes some costs. Field data on population-specific differences in premigratory fuel stores in songbirds are scanty. We were able to highlight the existence of such differences and to relate them to the overall width of the ecological barriers to be crossed when leaving the pre-migratory fuelling areas. A more precise identification of the geographical origin of birds building up energy stores in Spain and Italy, which will hopefully be provided by further developments of the EURING Swallow Project, together with a detailed quantification of the physiological costs of fattening and migration in the populations studied (e.g. Drent and Piersma 1990, Biebach 1998, Piersma 1998), will be a further step in the research into the migration ecology of this species. Acknowledgements – We wish to acknowledge all volunteer ringers who enthusiastically took part in the field activities of the EURING Swallow Project: this work would have been impossible without their invaluable support. G. Bogliani (University of Pavia) provided stimulating discussion and input into the ideas presented in this paper. Useful comments by T. Piersma and two reviewers improved the conceptual clarity of this paper.

References Bairlein, F. 1985. Body weights and fat deposition of Palaearctic passerine migrants in the central Sahara. – Oecologia 66: 141 – 146. Bairlein, F. 1988. Herbstlicher Durchzug, Ko¨ rpergewichte und Fettdeposition von Zugvo¨ geln in einem Rastgebiet in Nordalgerien. – Vogelwarte 34: 237 – 248.

21

Bairlein, F. 1991. Body mass of Garden Warblers (Syl6ia borin) on migration: a review of field data. – Vogelwarte 36: 48 – 61. Bairlein, F. 1992. Recent prospects on trans-Saharan migration of songbirds. – Ibis 134 (suppl.): 41 –46. Berthold, P. 1975. Migration: control and metabolic physiology. – In: Farner, D. S. and King, J. R. (eds). Avian Biology. Vol. 5. Academic Press, London, pp. 77 – 128. Berthold, P. 1993. Bird Migration. A general survey. – Oxford University Press, Oxford. Berthold, P. 1996. Control of Bird Migration. – Chapman & Hall, London. Berthold, P. and Friedrich, W. 1979. Die Federla¨ nge: eine neues nu¨ tzliches Flu¨ gelmass. – Vogelwarte 30: 11 – 21. Berthold, P., Helbig, A. J., Mohr, G. and Querner, U. 1992. Rapid microevolution of migratory behaviour in a wild bird species. – Nature 360: 368 –370. Biebach, H. 1990. Strategies of trans-Sahara migrants. – In: Gwinner, E. (ed.). Bird Migration: Physiology and Ecophysiology. Springer, Berlin, pp. 352 –367. Biebach, H. 1992. Flight range estimates for small trans-Sahara migrants. – Ibis 134 (suppl.): 47 –54. Biebach, H. 1998. Phenotypic organ flexibility in Garden Warblers Syl6ia borin during long-distance migration. – J. Avian Biol. 29: 529 –535. Biebach, H., Biebach, I., Friedrich, W., Heine, G., Partecke, J. and Schmidl, D. 2000. Strategies of passerine migration across the Mediterranean Sea and Sahara Desert: a radar study. – Ibis 142: 623 –634. Biebach, H., Friedrich, W., Heine, G., Jenni, L., Jenni-Eiermann, S. and Schmidl, D. 1991. The daily pattern of autumn bird migration in the northern Sahara. – Ibis 133: 414 – 422. Bibby, C. J. and Green, R. E. 1981. Autumn migration strategies of Reed and Sedge warblers. – Ornis Scand. 12: 1 – 12. Brown, M. E. 1996. Assessing body condition in birds. – In: Nolan Jr., V. and Ketterson, E. D. (eds). Current Ornithology. Vol. 13. Plenum Press, New York, pp. 67 – 135. Chilgren, J. D. 1977. Body composition of captive whitecrowned sparrows during postnuptial moult. – Auk 94: 677 – 688. Cramp, S. 1998. The Complete Birds of the Western Palearctic on CD-ROM. – Oxford University Press, Oxford. Davis, P. 1965. Recoveries of Swallows ringed in Britain and Ireland. – Bird Study 12: 151 –169. Drent, R. and Piersma, T. 1990. An exploration of the energetics of leap-frog migration in Arctic breeding waders. – In: Gwinner, E. (ed.). Bird Migration: Physiology and Ecophysiology. Springer, Berlin, pp. 399 –412. Jenni, L. 1998. EURING Swallow Project. Field manual (revised version). – Schweizerische Vogelwarte, Sempach. Jenni, L. and Winkler, R. 1989. The feather-length of small passerines: a measure for wing-length in live birds and museum skins. – Bird Study 36: 1 –15. Kaiser, A. 1992. Fat deposition and theoretical flight range of small autumn migrants in Southern Germany. – Bird Study 39: 96 –110. Kaiser, A. 1993. A new multi-category classification of subcutaneous fat deposits in songbirds. – J. Field Ornithol. 64: 246 – 255. King, J. R. and Murphy, M. E. 1985. Periods of nutritional stress in the annual cycle of endotherms: fact or fiction? – Am. Zool. 25: 955 –964. Lima, S. L. 1986. Predation risk and unpredictable feeding conditions: determinants of body mass in wintering birds. – Ecology 67: 377 –385. Lindstro¨ m, A, . and Piersma, T. 1993. Mass changes in migrating birds: the evidence for fat and protein storage re-examined. – Ibis 135: 70 –78.

22

McNamara, J. M., Houston, A. I. and Lima, S. L. 1994. Foraging routines of small birds in winter: a theoretical investigation. – J. Avian Biol. 25: 287 – 302. Mead, C. 1983. Bird Migration. – Country Life, Feltham. Moreau, R. E. 1961. Problems of Mediterranean-Saharan migration. – Ibis 103a: 373 – 427. Moreau, R. E. 1972. The Palaearctic-African Bird Migration Systems. – Academic Press, London. Moreau, R. E. and Dolp, R. M. 1970. Fat, water, weights and wing-lengths of autumn migrants in transit on the northwest coast of Egypt. – Ibis 112: 209 – 228. Murphy, M. E. 1996. Energetics and nutrition of molt. – In: Carey, C. (ed.). Avian Energetics and Nutritional Ecology. Chapman & Hall, New York, pp. 158 – 198. Newton, I. 1968. The temperatures, weights, and body components of molting bullfinches. – Condor 70: 323 – 332. Ormerod, S. J. 1989. The influence of weather on the body mass of migrating Swallows Hirundo rustica in South Wales. – Ring. Migr. 10: 65 – 74. Ormerod, S. J. 1991. Pre-migratory and migratory movements of Swallows Hirundo rustica in Britain and Ireland. – Bird Study 38: 170 – 178. Pennycuick, C. J. 1996. Wingbeat frequency of birds in steady cruising flight: new data and improved predictions. – J. Exp. Biol. 199: 1613 – 1618. Piersma, T. 1998. Phenotypic flexibility during migration: optimization of organ size contingent on the risks and rewards of fueling and flight? – J. Avian Biol. 29: 511 – 520. Pilastro, A. and Magnani, A. 1997. Weather conditions and fat accumulation dynamics in pre-migratory roosting Barn Swallows Hirundo rustica. – J. Avian Biol. 28: 338 – 344. Pilastro, A. and Spina, F. 1999. Fat accumulation in pre-migratory roosting Barn Swallows in Europe. – Acta Congr. Int. Orn. 22: 219 – 228. Redfern, C. P. F., Slough, A. E. J., Dean, B., Brice, J. L. and Hope Jones, P. 2000. Fat and body condition in migrating Redwings Turdus iliacus. – J. Avian Biol. 31: 197 – 205. Rice, W. R. 1989. Analyzing tables of statistical tests. – Evolution 43: 223 – 225. Rubolini, D. 2000. Strategie di migrazione della Rondine Hirundo rustica in Europa. – Tesi di laurea, Univ. of Pavia. Schaub, M. and Jenni, L. 2000a. Fuel deposition of three passerine bird species along the migration routes. – Oecologia 122: 306 – 317. Schaub, M. and Jenni, L. 2000b. Body mass of six long-distance migrant passerine species along the autumn migration route. – J. Orn. 141: 441 – 460. Spina, F. 1998. The EURING Swallow Project: results from the first pilot year. – In: Spina, F. (ed.). EURING Newsletter 2. INFS, Ozzano, pp. 22 – 27. Spina, F. and Massi, A. 1992. Post-nuptial moult and fat accumulation of the Ashy-headed Wagtail (Motacilla fla6a cinereocapilla) in Northern Italy. – Vogelwarte 36: 211 – 220. Svensson, L. 1992. Identification Guide to European Passerines. – Privately published, Stockholm. TCI 1998. Il nuovissimo Atlante del Touring. Vols. 1 – 2. – Touring Editore, Milano. Veiga, J. P. 1986. Settlement and fat accumulation by migrant Pied Flycatchers in Spain. – Ring. Migr. 7: 85 – 98. Witter, M. S. and Cuthill, I. E. 1993. The ecological costs of avian fat storage. – Phil. Trans. R. Soc. Lond. B 340: 73 – 90. (Recei6ed 8 January 2001, re6ised 28 May 2001, accepted 14 June 2001.)

JOURNAL OF AVIAN BIOLOGY 33:1 (2002)