Behavioral Ecology Advance Access originally published online on July 24, 2008
Behavioral Ecology 2008 19(6):1351-1360; doi:10.1093/beheco/arn066
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Trade-off between migration and reproduction: does a high workload affect body condition and reproductive state?
a Department of Biological Rhythms and Behaviour, Max Planck Institute for Ornithology, Von-der-Tann-Strasse 7, 82346 Andechs, Germany b Centre for Isotope Research, University of Groningen, Nijenborgh 4, 9747 AG Groningen, the Netherlands c Zoological Laboratory, University of Groningen, PO Box 14, 9750 AA Haren, the Netherlands d Max Planck Institute for Ornithology, Vogelwarte Radolfzell, Schlossalle 2, 78315 Radolfzell, Germany e Department of Behavioural Ecology and Evolutionary Genetics, Max Planck Institute for Ornithology, PO Box 1564, 82319 Seewiesen, Germany
Address correspondence to B. Kempenaers. E-mail: b.kempenaers{at}orn.mpg.de.
Received 25 October 2006; revised 15 May 2008; accepted 16 May 2008.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Migratory birds have to invest much energy into flight to reach their summer and winter quarters. Many studies have shown how migration affects body physiology, including the accumulation of energy stores and the reduction of nonessential organs. In spring, the costs of migration may trade-off with preparations for breeding, such as the timing and extent of development of primary and secondary sexual traits. Birds arriving earlier on the breeding grounds often have a higher reproductive success than late-arriving birds, but no study to date has addressed whether and how the flight workload during migration itself influences reproduction. Using a wind tunnel, we investigated the effect of a high workload during long flights on measures of body condition and reproductive state in male rose-colored starlings (Sturnus roseus). We compared an experimental group that flew in the wind tunnel every day and covered a total flight distance of >4700 km in 49 days with a control group of males that did not fly. All birds had ad lib access to food. After the "migration" period, individuals from both groups were kept in a common breeding aviary, where they directly competed for nest-boxes and females. Contrary to expectation, birds from the experimental and control group did not differ significantly in the spontaneous seasonal changes in fat score, in breast muscle thickness, in plasma testosterone levels, and in bill and mantle color. Body mass increased more slowly in experimental than in control birds, but it reached the same level soon after the migration period. We did not observe any effect of the experimentally increased heavy workload on behavior during the early breeding phase or on any parameter of reproductive success. We thus failed to find a trade-off between long flight and the development of traits in preparation for breeding or reproductive success. A possible treatment effect might have been obscured by the unrestricted food supply. However, we cannot exclude effects on other life-history stages such as future survival, migration, or reproduction. Our results attest to the strong endogenous control of seasonal physiological changes in preparation for breeding that occur independently of the extreme effort invested in long-distance migration.
Key words: body mass, energy expenditure, migration, reproduction, testosterone.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For many birds, life is particularly taxing during spring. Migratory birds have to invest much energy into flight during long-distance migration, which may lead to a trade-off with investment in reproduction immediately thereafter. The flight costs themselves are also subject to trade-offs, which have been modeled in optimal migration theories. Birds should either minimize the energy or the time spent during flight (Alerstam and Lindström 1990
; Lindström and Alerstam 1992; Hedenström and Alerstam 1997). Flying more economically may result in later arrival, but this also comes at a cost. Earlier arriving birds have a larger choice of breeding sites, territories, and mates (Marra et al. 1998) and a broader time window to recover from migration and to reproduce. Birds arriving earlier at the breeding grounds often occupy better territories and gain a higher pairing and fledging success (Alatalo et al. 1986; Lozano et al. 1996; Hasselquist 1998; Naef-Daenzer et al. 2001).
Migratory flights are demanding, and several studies have tracked changes in physiology and body composition during that time. Body mass and breast muscle (M. pectoralis) size decreased during long flights (Swaddle and Biewener 2000; Lindström et al. 2000; Schmidt-Wellenburg et al. 2008), and several organs are reduced during migration (Hume and Biebach 1996; Piersma and Lindström 1997; Biebach 1998; Battley et al. 2000). On the other hand, there is evidence that gonadal development already starts during spring migration in preparation for the upcoming breeding period (Bauchinger 2002; Bauchinger et al. 2005; Ramenofsky and Wingfield 2006), which should lead to a trade-off between investment in migration and in reproduction.
Although such a trade-off can be expected, no study to date investigated whether the high workload associated with prolonged flight affects subsequent reproduction. In zebra finches (Taeniopygia guttata), individuals with a higher workload (hopping activity) delayed reproduction, without affecting clutch size (Deerenberg and Overkamp 1999). Zebra finches also delayed reproduction when they experienced a reduced food intake rate and an increase in daily energy expenditure (DEE) during 6 weeks before reproduction but were kept under ad lib food conditions during breeding (Wiersma and Verhulst 2005). In these 2 studies, researchers manipulated workload, while simultaneously restricting access to food. Here, we decoupled the effects of an increased workload and caloric restrictions.
To understand how birds deal with having to invest in flight, prepare for reproduction, and optimally time their arrival, we need to assess the cost of migration in terms of reduced condition during and after migration and in terms of breeding performance. The aim of this study was to experimentally investigate the effect of a high workload associated with prolonged flight on an individual's condition and on its subsequent reproductive performance when food was available ad libitum. Several studies have shown that food restriction strongly affects body condition (Lindström et al. 2000; Battley et al. 2001), whereas we aimed to test the effect of the high workload per se. Thus, this study does not address all the costs associated with migration (e.g., food restriction and thermoregulatory costs; see Wikelski et al. 2003).
We used the migratory rose-colored starling (Sturnus roseus) as a model species because the birds easily breed in captivity and quickly learn to fly in a wind tunnel (Engel et al. 2006). During the spring migratory season, we let an experimental group of males fly about 4700 km each in a wind tunnel, which corresponds to their natural migration distance. A second group of males was kept as control under the same environmental and feeding conditions, but without the flights. During the "migration" period, we measured the DEE with the doubly labeled water (DLW) method (Speakman 1997). We also recorded changes in body mass, fat store, pectoralis size, levels of circulating testosterone, and bill and plumage coloration. After the migration period, all birds of both groups were brought into a large aviary where they could directly compete for nest sites and females. Under the hypothesis that prolonged flight during migration is costly and that there is a trade-off with investment in breeding, we predicted that the experimental group would show 1) a larger decrease in body mass and pectoral muscle size (Lindström et al. 2000; Swaddle and Biewener 2000); 2) delayed development of the gonads, resulting in lower plasma testosterone levels; 3) less intense coloration especially of the bill (presumably a sexual trait); and 4) delayed and less successful reproduction compared with the control group (Deerenberg and Overkamp 1999; Wiersma and Verhulst 2005). All predictions are summarized in Table 1.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Birds
Rose-colored starlings are migratory birds with their breeding range stretching from Central Asia to Eastern Europe and their wintering grounds from Pakistan to South India (Hudde 1993
). Birds live and migrate in flocks and breed in colonies. They build their nests in cliffs, under roofs, or in stone walls (Schenk 1929; Augustijn 1997; Miltschev and Tschobanov 2002). In the Ukraine, we observed that males arrived earlier in the breeding area than females, defended one or more nest sites, and guarded their mate. During spring, plumage coloration became more intense "rose" (personal observations) and bill color changed from a pale pink and brown to a more intense magenta and bluish-black (Hudde 1993). Male rose-colored starlings present their back to females when "dancing" during courtship, suggesting that the coloration of the mantle feathers may be a sexual signal. The birds in this experiment originated from a breeding colony on the Crimea peninsula, Ukraine. In 2001, we took 7-day-old nestlings and hand-raised them at the Max Planck Institute for Ornithology in Seewiesen, Germany. All experimental birds were used to handling, had already learned to fly in the wind tunnel, and were experienced breeders. We randomly assigned 8 males to the experimental group, which migrated in the wind tunnel, and 8 males to the control group.
Housing during the migration period
We kept the 16 males in groups of 4 in aviaries (ca. 2 m x 2 m x 2 m) adjacent to the wind tunnel. Birds of the experimental and the control groups were kept separated. Each aviary was equipped with 3 perches and 2 feeding trays. Birds were fed standard food (dried insects, heart, rusk, and eggshell, supplemented with minerals and vitamins), Realpasto, and fresh fruit ad libitum and had access to fresh water. We used Osram Biolux lamps simulating natural daylight. Day length was adjusted weekly according to the photoperiodic conditions the birds would have experienced during spring migration in the field, without accounting for the shift of photoperiod caused by the west-east migration. At the beginning of simulated migration on 1 March, the light:dark cycle (LD) was 12.4:11.6 (12.4 h from dawn civil twilight to dusk civil twilight) equivalent to Northern India (27.5 °N). By the end of migration on 18 April, day length had increased to LD 14.6:9.4, corresponding to the conditions in Southern Ukraine (47.5 °N, Figure 1A).
|
Migration in the wind tunnel
We let experimental birds fly in groups of 4 in our wind tunnel in Seewiesen (for a detailed description, see Engel et al. 2006
). Birds flew during a total period of 7 weeks from 1 March to 18 April, 2004. During the first week, all experimental birds flew only short flights of 0.25–1.5 h (0.6 ± 0.8 h per day) to get used to the wind tunnel (Figure 1B). From the second week onwards, flight time increased to 3 h (working day) or 1 h (Saturday/Sunday). Over the 7-week period, daily flight time averaged 2.2 ± 1.2 h per bird, with a total flight time of at least 109 h. With the flight speed set at 12 ms–1, the experimental birds thus covered a total distance of >4700 km (Figure 1B). This is equivalent to the distance the birds would have migrated between Northern India and the Crimea, Ukraine. Unfortunately, data on daily flight duration or on daily migratory distance covered are not available for rose-colored starlings. Birds of the control group experienced the exact same conditions as experimental birds (e.g., temperature, day length, level of human presence), except that they did not fly in the wind tunnel. We did not prevent the control birds from locomotory activities such as hopping or short flights within the aviaries.
Measurements
During the migration period, we measured physiological and morphological variables for all birds following the schedule indicated in Figure 1C.
DEE and body composition
We measured the energy expenditure of all individuals during 24 h (DEE) on 31 March, halfway through the migration period. Birds of the experimental group flew for 3 h in the wind tunnel during the measurement.
In the morning, we injected all 16 birds intraperitoneally with 0.22-g DLW, weighed on an analytical balance (Sartorius BP 1215) to the nearest 0.01 mg. Isotope enrichments for 18O and 2H were 60.4 and 36.5 atom percent, respectively. To let the DLW evenly equilibrate with the body water, we kept each bird in a dark box without access to food or water. After 1 h, we took an initial blood sample of about 60 µl from the jugular or brachial vein and weighed each bird to the nearest 0.01 g (initial body mass). We then brought each bird back to the aviary, with access to food and water. We started the flights at least 2 h after initial blood sampling to minimize effects of handling and bleeding on flight performance. We took the final 60-µl blood sample 24 h after the first sampling and weighed each bird again to the nearest 0.01 g (final body mass). We calculated the average body mass as the mean of initial and final body mass. We also collected blood samples from uninjected rose-colored starlings, which were not used in the experiment but kept under similar conditions, to assess the background levels of 2H and 18O. We subdivided all blood samples over 4 capillaries, immediately flame-sealed them, and stored them at 8 °C for further analysis in the Centre for Isotope Research, Groningen, the Netherlands. Analyses were done in duplicate for the initial and in triplicate for the final blood samples. For analytical procedures, see Engel et al. (2006).
We calculated the rate of carbon dioxide production using equation 7.17 from Speakman (1997), which was converted to the level of energy expenditure by using an energy equivalent of 27.3 kJ L–1 CO2 based on a protein-rich diet (Gessaman and Nagy 1988). We calculated the water efflux rates (rH2Oout, g day–1) during the 24-h measurement period using equation 4 of Nagy and Costa (1980), assuming that the proportion of the water flux lost by evaporation was 0.25 (Speakman 1997), and using a fractionation factor of 0.94 as recommended by Speakman (1997). We estimated water influx rates (rH2Oin, g day–1) from the change in total body water (TBW) during the measurement and the fractionation-adjusted water efflux rate rH2Oout. We calculated the volume of TBW based on the difference in 18O concentrations of the background and the initial samples after equilibration, correcting for an overestimation of 1.8% (Speakman et al. 2001). Based on the TBW measurements, we then calculated the mass of lean wet tissue (which is considered to contain about 73% water, Speakman 2001) and the fat content. To account for differences in body size, data on TBW and water flux are given relative to body mass (TBW% in %, and rH2O in g kg–1day–1).
Body measurements
We recorded body mass, fat score, and relative breast muscle thickness between 9:30 and 11:00 CET on a weekly basis during the migration period and finally on 20 April, immediately before the birds entered the breeding aviary (Figure 1C). We measured body mass to the nearest 0.01 g with a Sartorius BL 1500 S balance and visually assessed fat scores according to Kaiser (1993) with the fat scores for furcula and abdomen averaged. We measured the height of the breast muscle relative to the sternum and used the average of 3 measurements as "relative breast muscle thickness."
Plasma testosterone
We measured plasma testosterone levels as an indicator of reproductive state at the start, during, and after the experimental treatment (Figure 1C). We collected 200 µl blood samples from the jugular vein with a syringe and rinsed them through heparinized capillaries to prevent agglutination. After centrifugation (10 min at 14 000 rpm), we stored the plasma first at –21 °C and then at –70 °C until analysis.
We measured testosterone concentrations in the plasma samples with a radioimmunoassay (RIA) following Goymann et al. (2006). We used antiserum against testosterone from Esoterix Endocrinology (Calabasas Hills, CA; No. 73-125), standard steroids from Sigma-Aldrich (Munich, Germany), and labeled steroids from Perkin Elmer (Rodgau, Germany). All chemicals used were of analytical grade.
We added double-distilled H2O to aliquots of plasma (15–100 µl) to a total volume of 300 µl and equilibrated samples overnight at 4 °C with 
1500 dpm of 3H-testosterone (Perkin Elmer NET-553) to estimate hormone recovery. We then added 4 ml dichloromethane. After 4 h of equilibration, we separated the organic phase from the aqueous phase by freeze-decanting. After a second extraction with 2 ml of dichloromethane, the samples were dried under a stream of nitrogen gas at 39 °C. We resuspended the extracts in 300 µl phosphate-buffered saline with 1% gelatin (PBSG) and left them overnight at 4 °C to equilibrate. We then transferred aliquots (80 µl) of each fraction and sample to scintillation vials, mixed them with 4 ml scintillation fluid (Packard Ultima Gold), and counted them to an accuracy of 2–3% in a Beckman LS 6000 β-counter to estimate individual extraction recoveries, which were 89.4 ± 1.1% (average ± 95% CI). We stored the remainder at –40 °C until the RIA.
We set up standard curves by serial dilution of stock standard solutions with a concentration range of 0.4–200 pg of testosterone (in duplicate). We then added dilutions of the antiserum (1/200) to controls and to duplicates of the sample fractions (2 x 100 µl). After 30 min, we added 13 500 dpm of 3H-testosterone and incubated the samples for 20 h at 4 °C. We separated antibody-bound and free fractions of testosterone at 4 °C by adding 0.5 ml dextran-coated charcoal. After 14-min incubation with charcoal, we spun the samples (3600 g, 10 min, 4 °C) and decanted the supernatant into scintillation vials at 4 °C. We counted the samples after adding the scintillation liquid and letting them equilibrate for at least 4 h. We calculated standard curves and sample concentrations for the assay using Immunofit 3.0 (Beckman Inc, Fullerton, CA). All samples were analyzed within one assay. The intraassay coefficient of variation was 11.1%, and the lower detection limit was 0.7 pg per tube. Data were log-transformed for statistical analysis.
Plumage and bill color
At the start and at the end of the experimental treatment (Figure 1C), we recorded the reflectance spectra of the bill and of the feathers of the mantle and the rump from 300 to 700 nm, which encompasses the visual sensitivity of passerine birds (Hart 2001). We used an Avaspec 2048 spectroradiometer and an Avalight-DHS Deuterium Halogen light source (Avantes, Eerbeek, the Netherlands), connected through a bifurcated fiber optic probe. A cylindrical plastic tube was mounted on the optic probe to exclude ambient light and to standardize the measuring distance. We measured each sample 5 times with the probe held perpendicular to the sample. For the bill, we measured the pink area close to the nostril, 3 times on the left side and 2 times on the right. The feathers on the rump and mantle had a black rim, but we measured the pinkish part only.
We calculated the reflectance relative to a white standard (WS-2) with the software Spectrawin (Top Sensor Systems). For further analysis, we averaged the 5 spectra for each sample and summarized mean reflectance over 4.4-nm steps ("binned"; Grill and Rush 2000). We analyzed the reflectance spectra of bill, mantle, and rump using principal components (Cuthill et al. 1999). The first 3 principal components explained most of the observed variance (bill: 99.1%, mantle: 97.3%; rump: 98.1%). PC1 explained most of the variability in reflectance (bill: 94.6%, mantle: 54.7%, rump: 80.8%) and is usually interpreted as achromatic brightness (Cuthill et al. 1999). PC2 explained 2.9% of the variability in the bill and 33.5% in the mantle spectra and corresponded to high reflectance in UV and red. For the rump, it explained 14.5% and corresponded to high reflectance of intermediate wavelengths. PC3 corresponded to high reflectance in UV in all analyses, explaining 1.5%, 9.0%, and 2.8% of the observed variability in reflectance for bill, mantle, and rump, respectively.
At the start and at the end of the experimental treatment, we also took digital pictures from the bill with a Canon Eos 10 D camera with a circular flash (MR 14EX) under standardized conditions (distance between the camera and the bill: 69 cm, focal width: F 22, exposure: 1/200 s, gray background). We adjusted white balance with the program "Capture One" and analyzed pictures with Photoshop to determine hue, saturation, and brightness of the pink area in the front of the left nostril. Each picture was measured 5 times. Hue is on a degree scale (0–360°), and our measurements lie between 340° and 5° (360° is red; Kuehni 1999; Harold 2001). We linearized the scale by setting 360 to 0°, so that values between 340° and 359° are negative (e.g., 356° became –4°).
The breeding aviary
The breeding aviary was 16.6 m long, 3.2 m wide, and 2.6 m high and was semiopen to let in daylight. We kept the birds under natural local photoperiodic conditions (47.5 °N), corresponding to those at their natural breeding grounds in the Ukraine. The aviary contained 8 equal sections (2.1 m x 3.2 m x 2.6 m), each with an artificial wall with 2 nest-boxes (0.21 m x 0.25 m x 0.21 m, with entrance hole diameter 5.3 cm). The nest-boxes were fixed behind a brick wall to resemble natural breeding sites. In each section, we provided straw, hay, fresh turf-sward, and coconut fibers as nesting material. We also supplied each section daily with the same amount of standard food (see above), fresh fruit and salad, and water. Thus, resources such as nest sites, food, and nesting material were distributed equally over the aviary.
We transferred all males from both treatment groups (N = 15, 1 experimental bird was excluded because of an injured leg) to the breeding aviary at 12:00 on 20 April (henceforth referred to as day 0). Immediately afterwards, we released an equal number of females (N = 15) into the aviary. Birds started inspecting nest-boxes on day 0 and started nest building on day 2.
Behavioral observations
We started observations 75 min after the birds had been transferred to the breeding aviary. Each section was observed for 15 min and again for 15 min later in the afternoon on day 0. Two people simultaneously observed the birds’ behavior in 2 sections. From days 1 to 3, each section was observed for 1 h (30 min starting at 08:00 h, 15 min starting at 10:30, and 15 min starting at 15:00). On days 4 and 5, each section was observed for 30 min, starting at 8:00. In total, each individual was observed during days 0–5 on average for 2.2 ± 0.6 h (range: 1.3–3.3 h). The total observation duration and the timing of the observations did not differ between the experimental and control groups.
We recorded the duration of singing and of other courtship displays (e.g., crest-raising, tail-fanning, wing-waving or flapping, presenting the back while singing, "dancing" around the female, presentation of nesting material) and the frequency of aggressive behaviors between males (threat display, chasing, physical fights). We also noted who won the interaction.
For the analyses, we calculated the proportion of time each behavior was observed, relative to the total observation time. For each male involved in an aggressive interaction, we also calculated the proportion of interactions won. Proportions were arcsine-transformed before analysis.
Recording of reproduction
Every second day after the morning behavioral observations, we checked the nest-boxes and recorded nest building, the number of eggs laid, and hatching. We do not have information about extrapair paternity; hence, estimates of male reproductive success refer to within-pair success only. We stopped the experiment after the first chick had hatched, that is, we did not consider clutches initiated after day 48.
Statistical analyses
We used SPSS 14.0 for all analyses and report 2-tailed tests. Data shown are average ± standard deviation. We used t-tests with separate variance estimates to analyze differences in the changes of parameters between the control and experimental groups and paired t-tests to analyze within-individual changes over time. We applied restricted maximum likelihood (REML) mixed models to test for changes over time if a parameter was measured at least 3 times during the migration period. Individual ID was entered in the model as a random factor, and time, group, and group x time were introduced as fixed factors. Models were simplified in a stepwise manner, starting with the exclusion of nonsignificant (P > 0.10) interactions.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Part 1: Migration period
Daily energy expenditure
DEE was on average 55% higher in experimental than in control birds (experimental (N = 8): 224.4 ± 21.0 kJ d–1; control (N = 8): 144.5 ± 25.0 kJ d–1; t = 6.94, P < 0.001). Correspondingly, both water influx (rH2Oin; experimental: 24.1 ± 3.7 g, control: 16.8 ± 3.4 g) and water efflux (rH2Oout; experimental: 24.6 ± 3.5 g, control: 17.6 ± 3.1 g) were about 40% higher in experimental than in control birds (rH2Oin: t = 4.14, P < 0.01; rH2Oout: t = 4.28, P < 0.01).
During the DEE measurements, body mass did not significantly differ between the 2 groups (experimental: 82.5 ± 9.1 g, control: 88.4 ± 11.1 g; t = 1.16, P = 0.27).
Body condition
During the first week of "migration," when birds were getting used to the daily handling routine, body mass decreased significantly (paired t = 4.28, N = 16, P < 0.01, Figure 2A). This effect was similar for experimental and control birds (t13.56 = 0.61, P = 0.55). Then, between 9 March and 13 April, body mass increased significantly (Figure 2A, REML mixed model; time: F5,40 = 16.41, P < 0.001). As predicted, this increase appeared to be smaller in experimental than in control birds (Figure 2A; group: F1,14 = 0.63, P = 0.44; group x time: F5,40 = 2.45, P = 0.05). As a consequence, body mass was lower in experimental than in control birds between 16 March and 13 April, when experimental birds experienced the high daily workload (group: F1,71 = 4.63, P < 0.05; time: F4,28 = 2.54, P = 0.06).
|
Similarly, fat scores increased significantly from 9 March onwards (Figure 2B; REML mixed model; time: F5,53 = 4.66, P < 0.01) with experimental birds showing significantly lower fat scores (group: F1,18.2 = 6.21, P < 0.05; the nonsignificant interaction group x time was excluded from the model).
Relative breast muscle thickness also increased significantly from 9 March onwards (Figure 2C; REML mixed model; time: F5,45 = 6.53, P < 0.001), but we did not observe a difference between the 2 groups (group: F1,18 = 0.62, P = 0.44; group x time: excluded from the model).
TBW in relation to body mass (TBW%) was higher in experimental (59.3 ± 5.1%) than in control birds (52.8 ± 4.5%; t13.79 = 2.69, P < 0.05) on 31 March. Experimental birds therefore had a higher lean wet mass (66.8 ± 3.05 g) and a lower amount of fat (16.1 ± 7.8 g) than control birds (lean wet mass: 63.9 ± 3.1 g; fat: 25.3 ± 8.9 g).
Plasma testosterone
Plasma testosterone levels increased during the migration period, but without a significant difference between the 2 treatments (Figure 3; REML mixed model with log-transformed data; time: F2,44 = 80.78, P < 0.001, group: F1,44 = 2.31, P = 0.14).
|
At the beginning of the migration period, plasma testosterone levels were 50.6 ± 10.0 pg ml–1 (range 38.2–62.3 pg ml–1) in the experimental and 48.9 ± 23.5 pg ml–1 (range 23.0–85.0 pg ml–1) in the control group, respectively. After 2 weeks, levels increased to 117.2 ± 140.5 pg ml–1 (range 43.3–455.2 pg ml–1) in experimental and to 70.2 ± 32.4 pg ml–1 (range 38.0–140.2 pg ml–1) in control birds.
At the end of the migration period, plasma testosterone levels were on average more than 70% higher in birds of the experimental (1412.6 ± 1091.5 pg ml–1; range 241.4–3082.9 pg ml–1) than in those of the control group (816.7 ± 684.3 pg ml–1; range 191.2–1862.1 pg ml–1), but the difference was statistically not significant, due to large individual variation combined with relatively small sample size (t13.94 = 1.20, N = 16, P = 0.25).
Coloration
The reflectance spectra of the bill show a peak in the UV at around 370 nm and an increase in reflectance from 600 nm onwards (Figure 4A). The principal component analysis of the reflectance spectra revealed a change of PC1 and PC2 with time (PC1: paired t = 2.27, P < 0.05, PC2: paired t = 6.02, P < 0.001, PC3: paired t = 1.10, P = 0.29; N = 16). This reflects a decrease in overall reflectance (PC1) and a relatively smaller decrease in the UV part of the spectrum and at longer wavelengths compared with other wavelengths (PC2). The change was independent of the treatment (Table 2).
|
|
The analysis of photographs of the bill revealed a significant decrease in hue, that is, a shift toward a more purplish color (paired t = 4.40, N = 16, P < 0.01); an increase in saturation, that is, richer colors (paired t = -5.35, P < 0.01); and a decrease in brightness (paired t = 2.26, P < 0.05; Table 2). None of the changes differed between the 2 groups (Table 2).
The reflectance spectra of the mantle and the rump showed high reflectance in the UV and at long wavelengths with a trough in between (Figure 4B,C). We did not detect a change in coloration in the mantle feathers, except for an increase in the reflectance at longer wavelengths, that is, feathers appeared pinker at the end of the migration period. Again, there were no differences between the groups (PC analyses, data not shown).
Condition upon arrival in the breeding aviary
On 20 April, when birds were transferred to the breeding aviary and 2 days after the last day of the migration period, the difference in body mass between the control and experimental birds had disappeared (Figure 2;t11.45 = 0.15, Nexperimental = 7, Ncontrol = 8, P = 0.88). There was also no difference in fat score and relative breast muscle thickness (Figure 2; t11.30 = 1.65, P = 0.13, and t11.97 = 0.59, P = 0.57, respectively).
Part 2: Breeding period
Behavior during day 0–5
Song activity accounted for 17.5 ± 4.5% of the time budget of the males during this period and courtship behavior for 13.4 ± 4.6% (Figure 5A,B). The highest song activity was observed between days 1 and 3, whereas birds were engaged in other courtship activities mainly between days 3 and 5. There was no difference between the experimental and the control groups in either song activity (t11.91 = 0.03, P = 0.98) or in courtship behavior (t11.38 = 0.13, P = 0.90).
|
Similarly, neither the frequency of aggressive male–male interactions nor the proportion of interactions won differed significantly between control and experimental birds (Figure 5C,D; frequency: t12.05 = 1.78, P = 0.09, interactions won: t9.17 = 0.73, P = 0.49).
Reproductive success
In total, only 8 of the 15 males bred successfully (i.e., paired and produced at least 1 clutch), 3 experimental and 5 control birds. Two clutches were produced for which no social male could be identified. One experimental and two control males were polygynous, that is, they paired with 2 females simultaneously.
There was no obvious difference in the timing of pairing between the control and experimental birds. Completed nests were observed on day 6.3 ± 2.9 in experimental and on day 7.8 ± 1.9 in control birds (Nexperimental = 3, Ncontrol = 5). Clutches were initiated on day 17.7 ± 3.2 in experimental and on day 20.6 ± 4.7 in control birds. Average clutch size was 8.0 ± 0.0 in experimental and 6.6 ± 0.8 in control birds. All 8 males fed the chicks. Given the low sample sizes, no formal statistical tests were conducted.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
During the migration period, we observed seasonal changes in morphological and physiological parameters. In both the migrating and the control groups, body mass, fat stores, relative breast muscle thickness, and plasma testosterone levels increased. All birds also showed a marked shift in coloration of bill and mantle feathers toward more pronounced UV and red. These changes were obviously endogenously triggered, probably as a direct consequence of increased day length (Gwinner 1973
; Gwinner 2003). However, we found surprisingly little evidence that the migratory flight affected body condition and subsequent breeding performance (Table 1).
During the migration period, experimental birds showed a nonsignificant slower increase in body mass, a significantly lower body mass during the time of intensive migration (16 March–13 April), and significantly lower fat scores compared with control birds. Because the energetic costs of flight are positively correlated with body mass, experimental birds might have kept body mass lower to save costs (Swaddle and Biewener 2000). In accordance, immediately after the migration period, experimental birds increased their body mass. By the time the birds were released into the breeding aviary, the 2 treatment groups did not significantly differ in any of the measured parameters.
In the breeding aviary, the singing, courtship, and aggressive performance of males from both groups were indistinguishable. Although we did not apply statistical tests due to the low sample size, the treatment groups did not obviously or consistently differ in reproductive success (if anything, the experimental birds did better).
Workload manipulation
Each experimental bird that flew for 3 h spent on average 80 kJ d–1 or about 55% more energy than a control bird. Extrapolating the additional costs of flight to the entire duration of migration results in an additional energetic cost of 2910 kJ for the 4700-km flight. This is as much energy as contained in 75 g fat and corresponds with almost the total body mass of a rose-colored starling at the beginning of the migration period. It also corresponds with the total amount of energy a control bird would spend during 20 days in the aviary.
Although experimental birds spent more energy, they did not significantly differ from control birds at the end of migration except in body mass and fat score. We now discuss 3 potential explanations for a lack of an effect.
1. Were migrating birds able to compensate for the increased workload?
Engel et al. (2006) estimated the flight costs for a 82.5-g rose-colored starling as 30.7 kJ h–1 (based on DLW measurements during prolonged flights in the wind tunnel) or 92.2 kJ during a 3-h flight. Our measurements show that the DEE was only 79.9 kJ higher in the experimental group than in the control group. The difference of 12.3 kJ may be attributable to the energy that control birds spent during 3 h of sitting in the aviaries. However, this would only amount to 4.1 kJ h–1, whereas we obtain a figure of 6.0 kJ h–1 (measured DEE of the control birds expressed in kJ h–1). The latter value is about 50% higher than the former, even without taking into account reduced energy expenditure at night (which would yield an even higher energy expenditure during daytime). In conclusion, the energy expenditure during nonflight must have been lower in experimental than in control birds.
Birds of the experimental group might have been less active in the aviaries when they were not flying and/or they might have had a lower nocturnal energy expenditure, as observed in European starlings (Sturnus vulgaris; Bautista et al. 1998) and zebra finches (Deerenberg et al. 1998). In the latter study, birds that were exposed to an increased workload (hopping between perches to obtain food) reduced their nocturnal energy expenditure to such an extent that their total DEE did not exceed that of control birds. In the European starlings, birds under the hard working regime reduced their food intake, body mass, and overnight metabolism, so that they even had a lower DEE than birds under the "easy" working regime (Bautista et al. 1998). In contrast, nocturnal energy savings could not compensate for hard work in a field study on great tits (Parus major) (Wiersma and Tinbergen 2003): DEE and feeding rates were higher in females with an experimentally increased brood size than in females with a reduced brood size. Nocturnal resting metabolic rate, however, was not affected by the treatment (Wiersma and Tinbergen 2003).
We provided all birds with ad libitum food to decouple effects of high workload and food restriction. Hence, birds of the experimental group may have been able to sufficiently refuel and thus compensate for potential negative effects of the energy expenditure during flight. During natural migration, food is probably more limited. Our data thus suggest that body condition and subsequent reproduction are not negatively influenced by the high workload during migration per se, as long as birds have access to stopover sites with abundant food.
2. Was the experimentally applied workload too low to affect the birds’ condition?
Although the experimental birds flew over a distance that corresponds to the distance they cover in nature, our imposed flight schedule may not match natural migration in important ways. Birds migrating under natural field conditions 1) will probably fly for longer periods of time per day with stopover periods in between, 2) will be food-restricted, and 3) will have to spend more energy during nonflight than our captive birds. Indeed, a study on New World Catharus thrushes showed that only about 30% of the total energy expenditure during migration is spent on flight, whereas about 70% is related to stopover, particularly to thermoregulatory costs during stopover (Wikelski et al. 2003).
We measured DEE in 9 free-living, male rose-colored starlings on the Crimea peninsula with the same method during the early breeding phase (May–June 2003; Schmidt-Wellenburg CA, unpublished data). Compared with these birds (henceforth referred to as "Ukraine," N = 9), DEE in experimental birds was slightly but not significantly higher (Ukraine: 199.0 ± 29.9 kJ d–1; t14.31= 2.04, P = 0.06). However, the Ukraine birds were lighter (73.9 ± 2.6 g) than the experimental birds (t8.01 = 2.61, P < 0.05), and DEE expressed relative to body mass did not differ between both groups (Ukraine: 2.69 ± 0.37 kJ g–1 d–1; experimental: 2.74 ± 0.32 kJ g–1d–1; t15 = 0.11, P = 0.91). The early breeding period is demanding for males because they spend most of their time fighting and in courtship displays. DEE during migration in the field may be higher than during the early breeding phase, but no data are available to test this.
3. Could birds compensate for the high workload in the short-term, but suffer long-term costs?
It is possible that the costs of migration were not expressed in the short-term fitness components considered here but would turn up if lifetime reproductive success would be measured. For example, a high workload may affect the probability of future survival.
Not much is known about long-term effects of high workloads. Kestrels (Falco tinnunculus) raising enlarged families with increased work rates and higher DEE died sooner than those raising reduced families (Daan et al. 1996). Two studies on zebra finches showed that reproduction was delayed in birds under a high workload compared with a lower workload (hopping for food), whereas there was no difference in clutch size (Deerenberg and Overkamp 1999; Wiersma and Verhulst 2005). Zebra finches under a high workload also replaced experimentally removed feathers with shorter ones, and the additional cost of feather replacement resulted in a further delay of reproduction (Wiersma and Verhulst 2005).
Several studies on chick development have shown carry-over effects into later life-history stages. Chicks reared in adverse conditions may be able to catch up until fledging (Metcalfe and Monaghan 2001; Fitze et al. 2004), but they often pay later in life in terms of reduced survival, lower social rank, worse territory quality, smaller clutch size, and ultimately, lower reproductive success (Gebhardt-Henrich and Richner 1998; Lindström 1999; Metcalfe and Monaghan 2001; Fitze et al. 2004). Gustafsson and Pärt (1990) showed that reproductive effort earlier in life affected clutch size in collared flycatchers (Ficedula albicollis), and they suggested that costly reproduction enhanced the rate of senescence. We cannot exclude that our experimental rose-colored starlings would pay later for the high workload experienced during the migration period compared with the control group.
Condition during the migration period
The increase in body mass, breast muscle thickness, and fat score during the migration period cannot be explained by food supply alone because birds were kept under the same food conditions before the experiment. Apparently, these changes are endogenously programmed. Birds may store extra resources anticipating the energy needs during migration, the unpredictable situation at the breeding grounds, or the high energetic costs particularly during the early breeding phase (Bromley and Jarvis 1993).
Similarly, changes in plasma testosterone levels and in bill and plumage color appeared to follow a seasonal program. These endogenously triggered changes took place even in the absence of females or nesting sites, which—in the field—might act as additional stimuli.
Lindström et al. (2000) showed rapid changes in both muscle size and body mass in red knots (Calidris canutus) during fasting, during flights of several hours in a wind tunnel, and during refueling. They hypothesized that changes in the pectoral muscle might be caused by protein metabolism, the need for a special protein:fat ratio, or by the regulation of flight capacity (a higher body mass requires a bigger muscle). Similarly, exercising European starlings (Sturnus vulgaris) showed a decrease in both body mass and pectoralis mass (Swaddle and Biewener 2000). We observed an effect of exercise on body mass but not on relative breast muscle thickness. Although we failed to measure an effect of migration on relative breast muscle thickness, our measurements suggest that birds of the experimental group had more muscle tissue, as their lean wet mass was higher than that of control birds.
A recent study suggests that an individual's condition might be less affected by migratory flight than previously assumed. Hasselquist et al. (2007) let red knots fly 15 000 km in a wind tunnel during 6 days and demonstrated that the immune response was not impaired in flying compared with nonflying birds. Only birds that refused to fly for prolonged periods of time showed a lower response. The authors concluded that red knots are well adapted to the high workload during migratory flights, and they suggested that only birds in good condition would be willing to perform migratory flights.
Breeding and reproduction
Experimental and control birds did not differ in activity, behavior, or performance in male–male interactions. Plasma testosterone levels also indicated that birds of both groups were ready to reproduce. This is in contrast to the 2 studies on zebra finches mentioned above: both reported delayed reproduction in birds that had been exposed to a high workload (Deerenberg and Overkamp 1999; Wiersma and Verhulst 2005) However, in both studies, birds had to work (at lower or higher levels) for their food. Therefore, effects of an increased workload per se and of food restriction could not be separated.
|
ACKNOWLEDGEMENTS |
|---|
We thank the Azov Black Sea Ornithological Station in Melitopol, Ukraine, and especially Igor Belaschkov and Petro Gorlov for their support on the Crimea. We are indebted to Heidrun Bamberg, James Dale, Sophia Engel, Wolfgang Goymann, Barbara Helm, Gerhard Hofmann, Claudia Mettke-Hofmann, Ingrid Schwabl, and Monika Trappschuh for comments on experimental setup and analysis, and for practical and moral support. The study would not have been possible without the help of Iris Biebach, Sabine Dietrich, and Maria Lauterbach who helped flying the birds in the wind tunnel and observed them during the breeding part. We thank Serge Daan and 2 anonymous reviewers for constructive comments on the manuscript. The experiments were carried out in accordance with the German legislation on the protection of animals.
|
FOOTNOTES |
|---|
* Author is recently deceased.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Alatalo RV, Lundberg A, Glynn C. Female pied flycatchers choose territory quality and not male characteristics. Nature (1986) 323:152–153.[CrossRef][Web of Science]
Alerstam T, Lindström Å. Optimal bird migration: the relative importance of time, energy, and safety. In: Bird migration—Gwinner E, ed. (1990) Heidelberg, Germany: Springer. 331–351.
Augustijn G. Zuchterfahrungen mit einem Nomadenvogel. Der Rostenstar (Sturnus (Pastor) roseus Lin). Europäische Vogelwelt (1997) 2:13–21.
Battley PF, Dekinga A, Dietz MW, Piersma T, Tang S, Hulsman K. Basal metabolic rate declined during long-distance migratory flight in great knots. Condor (2001) 103:838–845.
Battley PF, Piersma T, Dietz MW, Tang S, Dekinga A, Hulsman K. Empirical evidence for differential organ reductions during trans-oceanic bird flight. Proc R Soc Lond B (2000) 267:191–195.[Medline]
Bauchinger U. Phenotypic flexibility of organs during long-distance migration in Garden Warblers (Sylvia borin): implications for migratory and reproductive periods [dissertation]. (2002) München, Germany: Technische Universität München. 166.
Bauchinger U, Wohlmann A, Biebach H. Flexible remodeling of organ size during spring migration of the garden warbler (Sylvia borin). Zoology (2005) 108:97–106.[Web of Science][Medline]
Bautista LM, Tinbergen J, Wiersma P, Kacelnik A. Optimal foraging and beyond: how starlings cope with changes in food availability. Am Nat (1998) 152:543–561.[CrossRef][Web of Science][Medline]
Biebach H. Phenotypic organ flexibility in Garden Warblers Sylvia borin during long-distance migration. J Avian Biol (1998) 29:529–535.[CrossRef]
Bromley RG, Jarvis RL. The energetics of migration and reproduction of Dusky Canada geese. Condor (1993) 95:193–210.
Cuthill IC, Bennett ATD, Partridge JC, Maier EJ. Plumage reflectance and the objective assessment of avian sexual dichromatism. Am Nat (1999) 160:183–200.[CrossRef]
Daan S, Deerenberg C, Dijkstra C. Increased daily work precipitates natural death in the kestrel. J Anim Ecol (1996) 65:539–544.[CrossRef]
Deerenberg C, Overkamp GJF. Hard work impinges on fitness: an experimental study with zebra finches. Anim Behav (1999) 58:173–179.[CrossRef][Web of Science][Medline]
Deerenberg C, Overkamp GJF, Visser GH, Daan S. Compensation in resting metabolism for experimentally increased activity. J Comp Physiol B (1998) 168:507–512.[CrossRef]
Engel S, Biebach H, Visser GH. Metabolic cost of avian flight in relation to flight velocity: a study in rose-coloured starlings (Sturnus roseus Linnaeus). J Comp Physiol B (2006) 176:415–427.[CrossRef][Medline]
Fitze PS, Clobert J, Richner H. Long-term life-history consequences of ectoparasite modulated growth and development. Ecology (2004) 85:2018–2026.[CrossRef][Web of Science]
Gebhardt-Henrich SG, Richner H. Causes of growth variation and its consequences for fitness. In: Avian growth and development: evolution within the altricial-precocial spectrum—Ricklefs RE, Starck M, eds. (1998) New York: Oxford University Press. 324–339.
Gessaman JA, Nagy KA. Energy metabolism: errors in gas-exchange conversion factors. Physiol Zool (1988) 61:507–513.
Goymann W, Trappschuh M, Jensen W, Schwabl I. Low ambient temperature increases food intake and dropping production, leading to incorrect estimates of hormone metabolite concentrations in European stonechats. Horm Behav (2006) 49:644–653.[CrossRef][Medline]
Grill CP, Rush VN. Analysing spectral data: comparison and application of two techniques. Biol J Linn Soc (2000) 69:121–138.[CrossRef][Web of Science]
Gustafsson L, Pärt T. Acceleration of senescence in the collared flycatcher Ficedula albicollis by reproductive costs. Nature (1990) 347:279–281.[CrossRef][Web of Science]
Gwinner E. Circannual rthythms in birds: their interaction with circadian rhythms and environmental photoperiod. J Reprod Fertil. (1973) (Suppl 19):51–65.
Gwinner E. Circannual rhythms in birds. Curr Opin Neurobiol (2003) 13:770–778.[CrossRef][Web of Science][Medline]
Harold RW. An Introduction to Appearance Analysis. SS Number 84A Reprint from GATFWorld, the magazine of the Graphic Arts Technical Foundation. (2001) Available from: http://www.color.org/ss84.pdf.
Hart NS. The visual ecology of avian photoreceptors. Prog Retin Eye Res (2001) 20:675–703.[CrossRef][Web of Science][Medline]
Hasselquist D. Polygyny in great reed warblers: a long-term study of factors contributing to male fitness. Ecology (1998) 79:2376–2390.[CrossRef][Web of Science]
Hasselquist D, Lindström Å, Jenni-Eiermann S, Koolhaas A, Piersma T. Long flights do not influence immune responses of a long-distance migrant bird: a wind-tunnel experiment. J Exp Biol (2007) 210:1123–1131.
Hedenström A, Alerstam T. Optimum fuel loads in migratory birds: distinguishing between time and energy minimization. J Theor Biol (1997) 189:227–234.[CrossRef][Web of Science][Medline]
Hudde H. Sturnus roseus—Rosenstar. In: Handbuch der Vögel Mitteleuropas—von Blotzheim UNG, ed. (1993) Wiesbaden, Germany: Aula. 2115–2144.
Hume ID, Biebach H. Digestive tract function in the long-distance migratory garden warbler Sylvia borin. J Comp Physiol B (1996) 166:388–395.[CrossRef]
Kaiser A. A new multi-category classification of subcutaneous fat deposits of songbirds. J Field Ornithol (1993) 64:246–255.
Kuehni RG. Calculation of CIELAB Hue difference adjustment factors from an ideal hue circle. Col Res Appl (1999) 24:292–294.[CrossRef]
Lindström Å, Alerstam T. Optimal fat loads in migrating birds: a test of the time-minimization hypothesis. Am Nat (1992) 140:477–491.[CrossRef][Web of Science]
Lindström Å, Kvist A, Piersma T, Dekinga A, Dietz MW. Avian pectoral muscle size rapidly tracks body mass changes during flight, fasting and fuelling. J Exp Biol (2000) 203:913–919.[Abstract]
Lindström J. Early development and fitness in birds and mammals. Trends Ecol Evol (1999) 14:343–348.[CrossRef][Medline]
Lozano GA, Perreault S, Lemon RE. Age, arrival date and reproductive success of male American redstarts Setophaga ruticilla. J Avian Biol (1996) 27:164–170.[CrossRef]
Marra PP, Hobson KA, Holmes RT. Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science (1998) 282:1884–1886.
Metcalfe NB, Monaghan P. Compensation for a bad start: grow now, pay later? Trends Ecol Evol (2001) 16:254–260.[CrossRef][Medline]
Miltschev B, Tschobanov D. Brutverluste und Nahrung des Rosenstares Sturnus roseus in Südost-Bulgarien im Jahr 2000. Vogelwelt (2002) 123:99–103.
Naef-Daenzer B, Widmer F, Nuber M. Differential post-fledging survival of great and coal tits in relation to their condition and fledging date. J Anim Ecol (2001) 70:730–738.[CrossRef]
Nagy KA, Costa DP. Water flux in animals: analysis of potential errors in the tritiated water method. Am J Physiol (1980) 238:R454–R465.[Web of Science][Medline]
Piersma T, Lindström Å. Rapid reversible changes in organ sizes as a component of adaptive behaviour. Trends Ecol Evol (1997) 12:134–138.[CrossRef]
Ramenofsky M, Wingfield JC. Behavioral and physiological conflicts in migrants: the transition between migration and breeding. J Ornithol (2006) 147:135–145.[CrossRef]
Schenk J. Die Brutinvasion des Rosenstares in Ungarn im Jahre 1925. In: Verhandlungen des VI Internationalen Ornithologischen Kongresses in Kopenhagen (1929) Berlin: Germany. 250–264.
Schmidt-Wellenburg C, Engel S, Visser GH. Energy expenditure during flight in relation to body mass: effects of natural increases in mass and artificial load in rose coloured starlings. J Comp Physiol B. (2008) doi 10.1007/s00360-008-0267-7.
Speakman JR. Doubly labelled water. In: Theory and practice (1997) London: Chapman and Hall.
Speakman JR. Body composition analysis of animals (2001) Cambridge: Cambridge University Press.
Speakman JR, Visser GH, Ward S, Król E. The isotope dilution method for the evaluation of body composition. In: Body composition analysis of animals: a handbook of non-destructive methods—Speakman JR, ed. (2001) Cambridge: Cambridge University Press. 56–98.
Swaddle JP, Biewener AA. Exercise and reduced muscle mass in starlings. Nature (2000) 406:585–586.[CrossRef][Web of Science][Medline]
Wiersma P, Tinbergen JM. No nocturnal energetic saving in response to hard work in free-living great tits. Neth J Zool (2003) 52:263–279.[CrossRef]
Wiersma P, Verhulst S. Effects of intake rate on energy expenditure, somatic repair and reproduction of zebra finches. J Exp Biol (2005) 208:4091–4098.
Wikelski M, Tarlow EM, Raim A, Diehl RH, Larkin RP, Visser GH. Costs of migration in free-flying songbirds. Nature (2003) 423:704.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||