Introduction

Every year, billions of birds migrate across areas where food and water are not available, with non-stop flights that lead to a rapid exhaustion of energy reserves. Thus, stopovers for refuelling are necessary to successfully continue migration. The duration of stopovers is influenced by a number of factors, including endogenous programmes, environmental parameters such as weather and food availability, and physiological condition of each individual at arrival (reviewed by Jenni and Schaub 2003). Theory predicts that migratory strategies and stopover behaviour is dependent on fat deposition (Alerstam and Lindström 1990; reviewed by Hedenström 2008), but evidence for a determinant role of body condition for stopover duration has been reported only in long-distance migratory passerines (Fusani and Gwinner 2004; Fusani et al. 2009; Goymann et al. 2010). To date, our knowledge about stopover behaviour of short-distance migrants is still scarce.

In captivity, the extent of migratory disposition of nocturnal migrants is estimated by the intensity of nocturnal restlessness, or Zugunruhe (Naumann 1795–1817). The duration and intensity of Zugunruhe is correlated with the duration and intensity of migration in free-living birds (Berthold 1973). In garden warblers (Sylvia borin), it was noted that leaner birds show a diurnal pattern of activity, whereas heavier birds show low diurnal activity and intense nocturnal activity, i.e. Zugunruhe (Bairlein 1985). In northern wheatears (Oenanthe oenanthe), birds with large fuel loads showed a greater degree of Zugunruhe the night after capture at a stopover site and accumulated less fuel over the subsequent 2 days than birds caught with small fuel loads (Eikenaar and Schläfke 2013). These observations match with the hypothesis that lean birds would spend more than 1 day at a stopover site to refuel and recover, whereas fat birds would resume migration in the night following arrival (Biebach 1985). This hypothesis was confirmed by laboratory studies in which a non-stop flight followed by stopover was simulated in fasting-and-refeeding experiments (Biebach 1985; Gwinner et al. 1988; Fusani and Gwinner 2004). In spring, the effects of the food regimen on nocturnal activity during the night following food reintroduction were highly correlated with the amount of fat reserves in blackcaps (Sylvia atricapilla). Results were later confirmed by field studies in which body condition was correlated with the extent of Zugunruhe in garden warblers, whitethroats (Sylvia communis) and whinchats (Saxicola rubetra; Fusani et al. 2009). In particular, we observed that birds with high fat reserves showed intense Zugunruhe at night and little activity during the day, whereas birds with little fat reserves showed the opposite pattern, suggesting that the physiological condition influences stopover duration in European long-distance migrants (Fusani et al. 2009). Furthermore, we confirmed these results in a whole-island telemetry study, in which we demonstrated the influence of fat reserves on real minimum stopover duration (Goymann et al. 2010).

Both long- and short-distance migrants are thought to migrate using a time minimization strategy (reviewed by Hedenström 2008). In spring, selection favours early arrival and onset of breeding (Kokko 1999). Several studies have suggested that the endogenous circannual mechanisms controlling migratory parameters are particularly well developed in long-distance migrants (Bairlein and Gwinner 1994). Across areas where refuelling stopover sites are available, endogenous factors should dominate the duration and length of migratory flights, whereas condition and extent of energy reserves are expected to determine the stopover duration before crossing an ecological barrier (Jenni and Schaub 2003). Similarly, body condition is likely to be crucial in determining stopover duration after crossing a barrier. In fact, when crossing ecological barriers, birds are usually unable to stop and refuel. Furthermore, some field data suggest that long-distance migrants deposit more fat before migration and also maintain fat reserves longer than short-distance migrants (Bairlein and Gwinner 1994). Therefore, energy reserves are expected to strongly determine the stopover behaviour in short-distance migrants. Previous studies on four different species of North American thrushes wintering in Central America or in the north of South America and crossing the Gulf of Mexico showed that lean birds were less active at night than fat birds just after overcoming the ecological barrier (Yong and Moore 1993), revealing the priority of birds without fat load to replenish energy reserves.

A few studies have demonstrated a simultaneous effect of body condition and weather factors on stopover duration in short-distance migrants (Morganti et al. 2011; Bulyuk and Tsvey 2013). In fact, when weather condition are favourable (strong tailwinds, weak crosswinds, low rainfall, high temperature, high pressure), fat black redstarts (Phoenicurus ochruros) are highly likely to depart from the stopover site in the evening of the day of arrival (Morganti et al. 2011), and although weather conditions alone do not determine the duration of stopover in robins (Erithacus rubecula), birds that stop for longer than one night tend to depart when the wind is favourable (Bulyuk and Tsvey 2013). We thus predict that short-distance migrants rest at the stopover site in relation to their physiological condition, similar to long-distance migrants caught at the same stopover site (Fusani et al. 2009). Here, we studied the relationship between physiological condition and the intensity of Zugunruhe of short-distance migrants after crossing the Mediterranean Sea. Birds were caught at a stopover site and placed in cloth cages to measure activity for 1 day, and were released the following morning.

Our results demonstrate that the physiological condition predicts the intensity of Zugunruhe in all three species of short-distance migrants.

Materials and methods

The study was conducted on Ponza (Italy), a small island in the Tyrrhenian Sea (40.855°N, 12.858°E) located along one of the main Mediterranean migratory routes. A ringing station has been active on the island since 2002. We used 51 black redstarts (Phoenicurus ochruros), 98 robins (Erithacus rubecula) and 17 stonechats (Saxicola torquata), caught with mist-nets during spring ringing operations in March and April 2011 and 2013. These three species are short-distance migratory passerines that winter in the Mediterranean coasts of North Africa. Data from capture–recapture studies suggest that in the spring, robins migrate to central-northern Europe, mainly to Sweden and Baltic coasts, black redstarts migrate to central Europe, mainly to Germany, and stonechats migrate to southern-central Europe, mainly to Italy, Austria and Hungary (Spina and Volponi 2008). Data from the Ponza ringing station show that the peak timing of passage (median Julian day) in the years 2007–2014 ranged from 82 to 97 in robins, 79 to 94 in black redstarts, and 73 to 81 in stonechats (Cardinale unpublished). Birds were caught in the morning between 7:30 and 11:30, subcutaneous fat was scored on a scale of 0–8 and the size of the pectoral muscles on a scale of 0–3, and body mass was measured to the nearest 0.10 g (Bairlein 1994). By 12:00, birds had been placed in individual cloth cages such that they were visually isolated from one another. Each cage was equipped with an activity infrared sensor to record locomotor activity. Birds were given only water ad libitum and were exposed to natural daylight conditions. All birds were released the following morning within 1 h after sunrise.

We calculated the number of times the infrared sensor was activated during each 2-min period. We then calculated the average activity during the intervals from 12:00 to sunset (day activity) and from sunset to sunrise (Zugunruhe) based on civil twilight times (Greenwich mean time +1). As the activity data were saved every day at 24:00, we calculated the average activity during the intervals from sunset to 24:00 (Zugunruhe during the first part of the night) and from 24:00 to sunrise (Zugunruhe during the second part of the night), and we calculated Zugunruhe as the mean value of the two previous variables. Finally, we transformed all the calculated values of activity (Zugunruhe during the first part of the night, Zugunruhe during the second part of the night and Zugunruhe) by taking the natural logarithms [ln(x + 1)]. To test for the effects of body condition on Zugunruhe, we first performed an analysis with the fat scores. We chose this approach because the fat score is a normalized index which is directly comparable across species, and previous studies have shown it to be a robust, consistent predictor of migratory disposition (Fusani et al. 2009). It was not possible to use general linear models (multiple regression) to study the relationship between fat score and Zugunruhe across species because of the study design, i.e. the species is a categorical variable with multiple levels. Therefore, we used a non-parametric approach. First, we performed a non-parametric Spearman correlation between fat score and Zugunruhe, including all species. The effect of the species on the relationship between fat and Zugunruhe was then analysed by means of the Quade test, a non-parametric analysis of covariance. We also performed a Pearson correlation test, separately for each species, between Zugunruhe and the index CONDITION, which was extracted by means of principal component analysis (PCA) from body mass, fat score and muscle score.

All statistical analyses were performed with R v. 3.2.0 using a significance level of α = 0.05.

Results

A global analysis across the three species, considered together, on a total of 167 individuals (51 black redstarts, 99 robins, 17 stonechats) yielded a significant positive correlation between Zugunruhe and fat score (r s = 0.263, n = 167, p = 0.001; Fig. 1). By applying a Quade test, we tested whether the correlation found in this global analysis was influenced by the factor species. The analysis showed no significant effect of the species on the relationship between Zugunruhe and fat (F 2,157 = 2,673, p = 0.072).

Fig. 1
figure 1

Log-transformed intensity of Zugunruhe (mean ± SE) plotted against fat score for black redstart (circles), European robin (triangles) and European stonechat (squares). Zugunruhe was significantly correlated with fat score, with no specific effect of the species. See text for statistical details

Full size image

For each species, we extracted a single PCA component from the variables fat score, muscle score and body mass, termed CONDITION. In black redstarts, CONDITION (eigenvalue 1.600) explained 53.0 % of the total variance, and its correlations with each variable were as follows: body mass 0.890; fat 0.900; and muscle −0.030. In robins, CONDITION (eigenvalue 1.950) explained 65.0 % of the variance, and its correlations were as follows: body mass 0.860; fat 0.830; and muscle 0.720. In stonechats, CONDITION (eigenvalue 1.690) explained 56.0 % of the variance, and its correlations were as follows: body mass 0.920; fat 0.920; and muscle 0.010.

In black redstarts, CONDITION was significantly positively correlated with the intensity of Zugunruhe; in particular, CONDITION was significantly positively correlated with Zugunruhe during the first part of the night, and was positively, but not significantly, correlated with Zugunruhe during the second part of the night (Table 1; Fig. 2a–c).

Table 1 Correlations (Pearson) between Zugunruhe, Zugunruhe during the first part of the night, Zugunruhe during the second part of the night, and diurnal activity and the factor CONDITION extracted from the variables body mass, fat score and muscle score by means of principal component analysis
Full size table
Fig. 2
figure 2

Scatterplots of log-transformed activity variables against the principal component CONDITION extracted from muscle score, fat score and body mass. For each species (top: black redstart, a–c; middle: European robin, d–f; bottom, European stonechat, g, h, i) we show Zugunruhe across the entire night (left, a, d, g), from sunset to midnight (middle, b, c, h) and midnight to sunrise (right, c, f, i). An asterisk (*) next to the panel letter indicates a significant correlation. See Table 1 for the statistics

Full size image

In robins, CONDITION was significantly positively correlated with the intensity of Zugunruhe; specifically, CONDITION was significantly positively correlated with Zugunruhe during both the first and the second parts of the night (Table 1; Fig. 2d–f).

Finally, in stonechats, the overall correlation between CONDITION and Zugunruhe was not significant; however, CONDITION was significantly positively correlated with Zugunruhe during the first part of the night, but not with Zugunruhe during the second part of the night (Table 1; Fig. 2g–i).

The amount of diurnal activity was not significantly correlated with CONDITION in any species (Table 1).

Discussion

Our results show that in black redstarts and European robins caught at a spring stopover site, body condition correlated with the intensity of Zugunruhe. In stonechats, the correlation between body condition and Zugunruhe was significant only for the first part of the night. In all species, Zugunruhe was more strongly correlated with body condition in the first part of the night, and the lack of correlation with Zugunruhe across the entire night in stonechats was probably due to the smaller sample size. Overall, our study confirms that condition is a good indicator of migratory tendency in these three short-distance migrants.

The results support the hypothesis that energy reserves carried by birds when arriving at stopover sites are good predictors of stopover duration (Bairlein 1985; Biebach 1985), a theory that we confirmed in previous studies on long-distance migrants (Fusani et al. 2009; Goymann et al. 2010). Thus, energy reserves appear to be a major determinant of stopover duration, independent of the migration strategy. Since the experimental approach and setup in the present study on short-distance migrants was identical to that in the previous study on long-distance migrants (Fusani et al. 2009), we are confident that the relationship between body condition and the degree of Zugunruhe reveals actual similarities in the control of migratory programmes between short- and long-distance migrants. One interesting difference is the possible contribution of the muscle to overall condition: in contrast to that observed in three species of long-distance migrants (Fusani et al. 2009), in black redstarts and stonechats, the factor CONDITION was not significantly influenced by the muscle score.

Our study shows that individuals of the three study species express more intense Zugunruhe in the first part of the night. Telemetric data on European robins have shown that the take-off time is shifted towards the beginning of the night in birds having higher fuel loads and during shorter nights (spring versus autumn), suggesting that departure time is dependent on fuel load and on night duration (Bolshakov et al. 2007). Moreover, data on caged birds have shown that migration is stimulated by darkness (Gwinner 1967), and have reported a quiescent aphasic phase in birds during sunset in preparation for the migration (Ramenofsky et al. 2003). In this study, we used individuals with fat scores of 2 or greater, thus excluding very lean birds, and we conducted the fieldwork during spring migration. Our results are in agreement with these observations, and confirm that individuals with a sufficient fuel load preferably perform their migratory flights in the first part of the night during spring migration.

Several external factors, such as weather condition and food availability, and internal factors, such as endogenous programmes, interact and thus influence the migratory programme and the decision of whether to stay at a stopover site or resume migration (reviewed by Jenni and Schaub 2003). Overall, spring migration is thought to depend on a time minimization strategy (reviewed by Hedenström 2008), as birds arriving earlier at the breeding grounds may obtain better territories and mate earlier, thus enhancing their reproductive success compared to birds arriving late (Kokko 1999). The results of the present study, however, suggest that migratory behaviour strongly depends on the body condition of the animals, and a time minimization strategy should not override the necessity for short-distance migrants in poor condition to regain energy reserves. Endogenous circannual mechanisms controlling migratory parameters are particularly well developed in long-distance migrants, leading to high temporal precision (Bairlein and Gwinner 1994; Helm et al. 2005). In fact, the timing of migration in long-distance migrants likely relies mainly on endogenous rhythms and photoperiod cues (Both and Visser 2001; Coppack and Both 2002). On the other hand, migratory schedules in short-distance migrants, which are characterized by shorter migration routes between wintering and breeding grounds, may be more flexible, and the total duration of migration may depend more on physiological condition and environmental factors. According to this hypothesis, our findings show that migratory behavior of short-distance migrants is finely modulated by energy condition. Moreover, we confirmed previous observations that lean birds were less active than fat birds at night just after overcoming the ecological barrier (Yong and Moore 1993).

A novel contribution of our study was the capacity of our approach to separate the influence of environmental factors from that of internal factors. Our study birds were kept indoor in cages and under similar conditions during the recording. Therefore, their behaviour was not affected by factors such as wind or cloud cover. Moreover, the capture and the experiments were conducted across the entire migratory season, and thus calendar day and weather conditions preceding capture had no systematic effect on the results. Although it is likely that these factors introduced variability in the sample, the large sample size allowed us to uncover clear relationships between condition and migratory behaviour.

Eikenaar and Bairlein (2013) reported that birds maintained in captivity under fasting conditions demonstrated a greater degree of Zugunruhe, illustrating that the motivation to migrate, quantified by Zugunruhe, is affected by changes in food availability. This result is in line with observational field studies in which migrants were more likely to depart from a stopover site when food was scarce than when it was plentiful (Bibby and Green 1981; Ottich and Dierschke 2003). Experiments have suggested that stopover duration is very short if the rate of fuel deposition is low or negative (Biebach 1985; Gwinner et al. 1985; Yong and Moore 1993). Previous studies on three species of trans-Saharan migratory passerines found that all birds departed in the evening of the arrival day from sites with low food availability, whereas lean birds stayed at least 1 day at sites with high food availability (Biebach et al. 1986). Preliminary data from our research station on Ponza suggest that this stopover site probably provides opportunities to regain fat reserves (Trez et al. unpublished), and hence it is likely that stopover duration is not influenced by the lack of food availability and that lean birds prolong stopover to replenish energy reserves.

A few studies have demonstrated the role of body condition in stopover duration among short-distance migrants, although the influence of environmental factors could not be excluded. In black redstarts, stopover duration is longer in individuals with low fat stores, and birds generally resume migration when weather conditions are favourable (Morganti et al. 2011). Moreover, in European robins, departure decisions are dependent on the simultaneous influence of both energy and weather factors (Bulyuk and Tsvey 2013).

In all three species, day activity was not significantly correlated with body condition. These results contrast with previous reports of higher diurnal activity among lean birds at a stopover site observed in the long-distance migrant, the garden warbler (Bairlein 1985), and in four species of North American thrushes that cross the Gulf of Mexico to reach their wintering areas (Yong and Moore 1993), which could reflect the simultaneous needs of foraging and minimizing exposure to predators. In robins making stopovers on the Courish Spit, the proportion of recaptured individuals is nearly fourfold lower in fat birds than in lean individuals (Titov 1999), suggesting a diurnal pattern of activity in lean birds. One possible explanation for the discrepancy between our results and those of previous works may lie in the pattern of migration and the frequency and distribution of the stopovers, leading to different strategies of fat accumulation through the migratory flight and different decisions during stopovers. Hence, diurnal activity should depend on the simultaneous effect of numerous factors—not only body condition, food availability and predation risk, but also the need to replenish energy reserves in relation to the stopover position along the migratory route.

In conclusion, our study revealed the influence of physiological condition on stopover duration in short-distance migrants, providing additional evidence of the importance of fuel resources in the decision of birds to resume migration or prolong stopover, and confirming that short- and long-distance migrants follow similar stopover strategies after crossing an ecological barrier.