Adaptive fault bar distribution in a long-distance migratory, aerial forager passerine?

Blackwell Science, LtdOxford, UKBIJBiological Journal of the Linnean Society0024-4066The Linnean Society of London, 2005? 2005 854 455461 Original Article FAULT BAR DISTRIBUTION AND FEATHER FUNCTION D. SERRANO and R. JOVANI Biological Journal of the Linnean Society, 2005, 85, 455 461. With 3 figures Adaptive fault bar distribution in a long-distance migratory, aerial forager passerine? DAVID SERRANO* and ROGER JOVANI Department of Applied Biology, Estación Biológica de Doñana (CSIC), Pabellón del Perú, Avda. María Luisa s/n, 41013 Seville, Spain Received 5 April 2004; accepted for publication 25 September 2004 Fault bars are translucent bands produced by stressful events during feather formation. They weaken feathers and increase their probability of breakage, and thus could compromise bird fitness by lowering flight performance. It has been recently suggested ( fault bar allocation hypothesis ) that birds could have evolved adaptive mechanisms for reducing fault bar load on the feathers with the highest function during flight. We tested this hypothesis by studying first-year individuals of the long-distance migratory, aerial forager barn swallow Hirundo rustica. We predicted that fault bars should be less abundant on the outermost wing and tail feathers, but more frequent on the tail than on the outermost wing feathers. Accordingly, we found that fault bars occurred more often in tertials than in primaries or secondaries. Tail feathers had fewer fault bars than tertials, but more than primaries. Within the tail, the distribution pattern of fault bars was W-shaped, with the highest fault bar load occurring on the streamers and on the two central feathers. Because streamers are the most important tail feathers for flight performance, this finding seems to contradict the fault bar allocation hypothesis. However, flight performance is much less sensitive to changes in the shape of the tail than of the wings, which could explain why evolutionary forces have not counteracted the increase of fault bars associated with feather elongation during the recent evolution of streamers in the tail of hirundines.. ADDITIONAL KEYWORDS: barn swallow fault bar allocation hypothesis feathers flight requirements Hirundo rustica stress bands. INTRODUCTION Fault bars constitute the most widespread type of abnormalities in the feathers of birds (Riddle, 1908). They are seen as translucent bands running approximately perpendicular to the rachis and are caused by barbules being thinner or completely absent. Among the proximal factors causing fault bars, nutritional stress (Slagsvold, 1982; Machmer et al., 1992), and handling stress (Negro, Bildstein & Bird, 1994) are the most commonly evoked. However, mechanisms promoting this kind of feather aberration are still poorly understood, perhaps because fault bars could be the common outcome of an array of stressful events when the feather is developing (Machmer et al., 1992). *Corresponding author. E-mail: serrano@ebd.csic.es In fact, it is known that fault bars can be induced by the administration of exogenous corticosteroids, so they are also known as stress bands by veterinarians (Ritchie, Harrison & Harrison, 1994). Regardless of the causal mechanisms, fault bars seem to be indicative of susceptibility to stress, and they have been proved to be correlated with quality and fitness components of individuals (Blanco & de la Puente, 2002; Bortolotti, Dawson & Murza, 2002). Apart from being indicators of general quality of individuals, fault bars could directly affect bird fitness. They constitute a weakness of the feather, and thereby increase the probability of feather breakage. In comparison to feathers lost accidentally, broken feathers are not immediately substituted and birds must wait until the next regular moult to replace them. Thus, birds with broken flight feathers can suffer a significant temporal increase in their wing load 455

456 D. SERRANO and R. JOVANI (i.e. body weight/wing area), which is a critical factor affecting flight performance (Pennycuick, 1989). This may produce an increase in the energetic demands of flight, compromising bird fitness as shown by experimental studies in which a reduction of wing area lowered breeding success (Mauck & Grubb, 1995; Velando, 2002). Moreover, flight performance is essential not only for reproduction, but also for migration (Lindström et al., 2000), foraging (Bautista et al., 1998) and escaping from predators (Witter & Cuthill, 1993). Thus, feather breakage resulting from fault bars could seriously affect many aspects of bird life history. These ideas, together with previous evidence suggesting that fault bars are more abundant on tail and body feathers than on flight wing feathers (e.g. Slagsvold, 1982; Machmer et al., 1992; Bortolotti et al., 2002), recently motivated Jovani & Blas (2004) to hypothesize that birds should have evolved adaptive strategies to minimize the fitness costs of faultbarring. One important prediction of the fault bar allocation hypothesis is that natural selection should have evolved mechanisms to reduce fault bar load on the feathers with the highest strength requirements and function during flight. Jovani & Blas (2004) tested this hypothesis in white storks Ciconia ciconia and found that fault bars in both nestlings and adults occurred in a non-random fashion, with those wing feathers functionally most important for flying (i.e. outermost wing feathers) having the lowest number of fault bars. In this study, we tested for the second time the fault bar allocation hypothesis using a small passerine, the barn swallow Hirundo rustica L. Barn swallows are appropriate study models because they are long-distance migrants and feed exclusively on flying insects, so they have high flight requirements. Flight requirements are lower in the innermost than in the outermost wing feathers, particularly during flapping (Corning & Biewener, 1998), so we expected the number of fault bars to decrease from the innermost feathers (i.e. tertials) towards the tip of the wing (i.e. primaries). Because the role of the tail on flight performance refers to stability, balance and turning (Thomas, 1997), the intensity of the induced drag supported by tail feathers is expected to be lower than that of outer wing feathers involved in flying activities like flapping. Thus, according to the fault bar allocation hypothesis, we expected fault bars to be less frequent on outer wing flight feathers than on tail feathers. Finally, the outermost tail feathers of barn swallows (the so-called streamers) have an important aerodynamic function (Norberg, 1994). They generate most of the tail lift force because they alone define its maximum continuous width, so they are particularly susceptible to breakage (Thomas, 1997). Accordingly, we expected barn swallows to have evolved mechanisms to reduce fault bars in the tail streamers. METHODS FIELD PROCEDURES From 27 August to 17 September 2003, we captured barn swallows near Zaragoza, north-eastern Spain, in an open farmland devoted to maize and alfalfa crops. Birds were tape-lured and captured with mist nets. Each bird was ringed with a numbered aluminium band, examined for fault bars, measured following standardized protocols, and released. The nine primaries (hereafter P), six secondaries (S) and three tertials (T) of the left wing, as well as the 12 rectrices (R), were inspected for the presence of fault bars and their number was quantified. The outermost primary was not inspected owing to its tiny size in this species. We studied fault bars only in juvenile (i.e. first-year) barn swallows for two reasons. First, in comparison to adults that moult their wing feathers sequentially, feather growth of juveniles in the nest is relatively synchronic, which allows us to avoid the confounding effect of differential stress conditions experienced during the growth of the different feathers. Second, juveniles have been found to have more fault bars than adults in a number of species (e.g. Slagsvold, 1982; Hawfield, 1986; Jovani & Blas, 2004), so they are expected to provide enough sample sizes for statistical comparisons. STATISTICAL ANALYSES Data were analysed mainly by using generalized linear modelling techniques (McCullagh & Nelder, 1983). Given that fault bars in different feathers of the same barn swallow could hardly be seen as independent events, we modelled variance covariance structures by using generalized linear mixed models (GLMMs) to avoid pseudo-replication. We implemented appropriate link functions and error structures in the SAS macro GLIMMIX (Littell et al., 1996). First, we analysed the probability of having fault bars in each group of feathers, i.e. P, S, T and R, with a binomial distribution of errors and a logistic link function. Because the number of feathers differed between groups, we used a ratio as the response variable, where number of feathers with fault bars was the numerator and whole number of feathers per group the binomial denominator. Then we analysed the number of fault bars per feather having at least one fault bar with a Poisson distribution of errors and a log link function. Feather group was fitted as a fixed explanatory variable and individual as a random factor in both analyses. Posthoc comparisons were performed using the Contrast statement of SAS. Moreover, we used Wilcoxon

FAULT BAR DISTRIBUTION AND FEATHER FUNCTION 457 matched-pairs signed-rank tests for analysing how consistent those general (i.e. population level) results found in GLMM analyses were at the individual level. All tests were two-tailed. RESULTS Fault bars were quantified in 172 barn swallows, of which 154 (89.5%) had a total of 1059 fault bars in the feathers examined. They produced breakage of feathers in nine individuals (5.2%), mainly in the tail. Fault bars showed a clear aggregated distribution among swallows, many birds showing a low number of fault bars on the wing and tail, while some of them having many fault bars (Fig. 1). Moreover, among swallows the number of fault bars was positively correlated between tail and wing feathers (GLM; F 1,152 = 68.32, P < 0.0001, Fig. 1). The proportion of feathers with fault bars differed among the four groups of feathers (F 3,513 = 157.04, P < 0.0001, Fig. 2 inset). Within the wing of swallows, the probability of having a fault bar decreased dramatically from the innermost to the outermost feathers (Fig. 2). It decreased dramatically from T to S (F 1,513 = 249.06, P < 0.0001), and slightly between S and P (F 1,513 = 5.94, P = 0.0151). This pattern held within birds (Table 1). The proportion of tail feathers with fault bars was intermediate between S and T, differing significantly from both groups of feathers (S vs. R: F 1,513 = 81.34, P < 0.0001; T vs. R: F 1,513 = 151.71, P < 0.0001). The number of fault bars per feather having fault bars also differed among groups of feathers (F 3,486 = 84.42, P < 0.0001, Fig. 3 inset). Within the wing, P and S had a similar number of fault bars (F 1,486 = 2.12, P = 0.146), but had fewer fault bars than T (P vs. T: F 1,486 = 66.03, P < 0.0001; S vs. T: Table 1. Wilcoxon matched-pairs signed-rank tests for examining the difference on the proportion and abundance (i.e. mean number) of fault bars occurring on feather types P S vs. T (see Fig. 2 for the position of the feathers on the wing) on the 131 barn swallows Hirundo rustica with fault bars on the wing. Data refer to the number of individual swallows within each category % with fault bars Abundance of fault bars P S < T 124 124 P S > T 7 7 Wilcoxon test Z = -10.222, P < 0.0001 Z = -10.222, P < 0.0001 Figure 1. Frequency distribution of fault bars in the wing and tail feathers of young barn swallows Hirundo rustica. Inset figure shows the relationship between the number of fault bars in the tail and the wing of each individual, together with the line best fitting the data.

458 D. SERRANO and R. JOVANI 80 P9 T3 0.4 % of feathers with fault bars 60 40 20 0 S1 P S T R P9 P8 P7 P6 P5 P4 P3 P2 P1 S1 S2 S3 S4 S5 S6 T1 T2 T3 Wing feathers 0.2 0.0 R2R3R4R5R6R6R5R4R3R2 Tail feathers Figure 2. Mean (+SE) percentage of wing and tail feathers having fault bars in young barn swallows Hirundo rustica. Inset figure summarizes the percentage in each group of feathers. Mean number of fault bars 2.0 1.5 1.0 0.5 0.0 P9 S1 P S T R P9 P8 P7 P6 P5 P4 P3 P2 P1 S1 S2 S3 S4 S5 S6 T1 T2 T3 Wing feathers T3 1.0 0.5 0.0 R6 R6 R2R3R4R5R6R6R5R4R3R2 Tail feathers Figure 3. Mean (+SE) number of fault bars on the wing and tail feathers of young barn swallows Hirundo rustica. Inset figure summarizes the mean number of fault bars in each group of feathers. Note that 18 birds without fault bars are not shown.

FAULT BAR DISTRIBUTION AND FEATHER FUNCTION 459 F 1,486 = 46.76, P < 0.0001, Fig. 3). This pattern held within birds (Table 1). R had fewer fault bars than P (F 1,486 = 4.30, P = 0.0385), but did not differ from S (F 1,486 = 0.05, P = 0.8229). The number of fault bars in T was significantly higher than in R (F 1,486 = 203.66, P < 0.0001). Within the tail, feathers differed in both the probability of having fault bars (F 11,1881 = 10.02, P < 0.0001) and the abundance of fault bars (F 11,1881 = 11.64, P < 0.0001), showing a W-shaped distribution pattern (Fig. 3). DISCUSSION Most juvenile barn swallows had fault bars in primaries, secondaries, tertials or rectrices, producing breakage of feathers in about 5% of the birds. It is worth noting that all birds were captured just before migrating to their African wintering quarters, where juvenile barn swallows are the last individuals moulting their flight feathers (Møller, 1994). Thus, the recorded feather breakage resulting from fault bars is a minimum estimate that could increase dramatically during advanced migration and winter. Barn swallows showed a non-random distribution pattern of fault bars. Propensity of fault-barring decreased from the innermost (i.e. tertials) towards the outermost (i.e. primaries) wing feathers. Within the tail, fault bars showed a W-shaped distribution pattern, being more abundant in the two innermost and the two outermost tail feathers. In this way, the tertials and the two central rectrices had the highest fault bar loads, despite being the smallest feathers inspected within the wing and tail, respectively, and thus those with a shorter time to develop fault bars (Jenni & Winkler, 1994). Within the tail, however, the longer exposure of longer feathers to fault-barring could explain the W-shaped distribution pattern when the two central rectrices are not considered (Fig. 3). Although a number of authors have found fault bars to be less common on wing than in tail feathers (Slagsvold, 1982; King & Murphy, 1984; Hawfield, 1986), even suggesting that plumage groups important for flight seemed to contain the fewest fault bars (Machmer et al., 1992), the fault bar allocation hypothesis had not been explicitly formalized until recent times (Jovani & Blas, 2004). Proximate factors causing fault bars in birds are much in dispute, but it is widely recognized that they are the result of stressful events during feather growth (Ritchie et al., 1994). In barn swallows, a tail-manipulation experiment showed that adult males with elongated tails developed a higher frequency of fault bars in their tail feathers in the subsequent year, suggesting that a stressful event (in this case, having an elongated tail) translated into an increase in the load of fault bars (Møller, 1989a). The fault bar allocation hypothesis proposes that even if stressors are difficult to avoid, natural selection will result in minimizing fault bars in the more functional feathers for flight performance (Jovani & Blas, 2004). Our results support the idea that juvenile barn swallows produce fault bars in such a way, such that they are rarer in primaries, secondaries and, to a lesser extent, on tail feathers. To our knowledge, this is the second time that this hypothesis has been explicitly tested and supported, providing further evidence that birds have evolved physiological adaptations to counteract the negative effect of stressors on flight performance through fault bar avoidance. Although our results are of paramount importance in understanding the implications of stressful conditions from an evolutionary perspective, the precise physiological mechanisms involved in differential fault bar allocation are not known. In the closely related welcome swallow H. neoxena, it has been showed that young maintain wing feather growth at the expense of body mass under poor nutritional conditions, suggesting a functional priority of maintaining wing development and symmetry (Ashton & Armstrong, 2002). This lean-priority strategy in the wing feathers could also have evolved to avoid the development of fault bars, but unfortunately there is insufficient information about feather development in other feather groups under stressful conditions. With respect to the tail, a plausible evolutionary scenario is that fault bar allocation gained importance with the evolution of flight, with mechanisms evolving to avoid fault bars in those tail feathers critical for flight performance. This is supported by empirical data provided by Machmer et al. (1992), who found that fault bars decreased from the innermost to the outermost tail feathers of nestling ospreys Pandion haliaetus. Thus, we expected fault bar loads to decrease from the innermost to the outermost tail feathers in bar swallows, the streamers having thus the lowest load. This is because streamers receive little support from neighbouring feathers (Thomas & Balmford, 1995), and have a high functional importance in flight performance (Norberg, 1994). Birds less capable of enduring stressful events therefore have a higher risk of streamer breakage, which may cause asymmetry in the tail. In aerial foragers such as barn swallows there are direct aerodynamic benefits from having symmetrical tails, particularly during slowspeed flight and manoeuvring flight (Thomas, 1993; Evans, Martins & Haley, 1994; Norberg, 1994). These advantages of symmetry could be especially important during extreme environmental conditions such as dry winters, in which overall resource abundance is low (Møller, 1989b). For instance, intense selection against asymmetric individuals has been documented for the related cliff swallow Petrochelidon pyrrhonota during

460 D. SERRANO and R. JOVANI spells of cold weather (Brown & Brown, 1998). Moreover, if streamer breakage increases the costs of foraging even in benign environments, young individuals could be constrained in developing long, attractive streamers for their first breeding attempt (Møller, 1994). We found that fault bars were abundant in the two central feathers despite being the smallest tail feathers. However, fault bars increased centrifugally in the remaining feathers, the streamers showing the highest abundance of fault bars. If allocating fault bars in the less functional feathers for flight performance is beneficial, why have barn swallows not evolved mechanisms to reduce their load in the outermost tail feathers? Forked tails and streamers in hirundines are recent traits that have evolved independently several times (Møller, 1994). Moreover, aerodynamic performance is thought to be much less sensitive to changes in the shape of the tail than of the wing, as shown by levels of fluctuating asymmetry being higher in tails (Thomas, 1997). In this way, the combined effect of the tail s relative aerodynamic insensitivity to modifications in shape together with the recent evolution of forked tails, could explain why evolutionary forces have not counteracted the increase of fault bars associated with feather elongation in the tails of barn swallows. The results presented here and those previously reported for white storks (Jovani & Blas, 2004), a phylogenetically distinct species with a very different life style, suggest that fault bar avoidance could be a widespread adaptive mechanism in flying birds. However, our results suggest that differential fault bar allocation could have evolved, or is maintained, only on those groups of feathers with the highest sensitivity to aerodynamic performance. In particular, fault bars in the external flight feathers of the wing could strongly penalize individuals during feeding and long-distance migration, but their effect on the tail feathers may not be so dramatic. In order to elucidate these postulations, we first need a detailed knowledge of the frequency and conditions of fault bars producing feather breakage. Moreover, it is vital to refine our understanding of how feather breakage in the different groups of feathers affects flight performance and fitness. Fault bar allocation could vary with ecological and evolutionary circumstances, so future studies should test this hypothesis in birds with different life histories such as non-migratory birds and ground foragers with low flight requirements. ACKNOWLEDGEMENTS This research was conducted under a ringing license from the Ministerio de Medio Ambiente of Spain. We are grateful to Francis Hernández for drawings and Julio Blas for stimulating comments. We especially thank Armando Serrano and Ángeles Larraz for their company and hospitality during the field work. REFERENCES Ashton JC, Armstrong DP. 2002. Facultative prioritization of wing growth in the Welcome swallow Hirundo neoxena. Ibis 144: 470 477. Bautista LM, Tinbergen J, Wiersma P, Kacelnik A. 1998. Optimal foraging and beyond: how starlings cope with changes in food availability. The American Naturalist 152: 543 561. Blanco G, de la Puente J. 2002. Multiple elements of the Black-billed magpie s tail correlate with variable honest information on quality in different age/sex classes. Animal Behaviour 63: 217 225. Bortolotti GR, Dawson RD, Murza GL. 2002. Stress during feather development predicts fitness potential. Journal of Animal Ecology 71: 333 342. Brown CR, Brown MB. 1998. Intense natural selection on body size and wing and tail asymmetry in cliff swallows during severe weather. Evolution 52: 1461 1475. Corning WA, Biewener AA. 1998. In vivo strains in pigeon flight feather shafts: implications for structural design. Journal of Experimental Biology 201: 3057 3066. Evans MR, Martins TLF, Haley M. 1994. The asymmetrical cost of tail elongation in red-billed streamertails. Proceedings of the Royal Society of London B 256: 97 103. Hawfield EJ. 1986. The number of fault bars in the feathers of Red-tailed Hawks, Red-shouldered Hawks, Broad-winged Hawks, and Barred Owls. The Chat 50: 15 18. Jenni L, Winkler R. 1994. Moult and ageing of European passerines. London: Academic Press. Jovani R, Blas J. 2004. Adaptive allocation of stress-induced deformities on the feathers of birds. Journal of Evolutionary Biology 17: 294 301. King JR, Murphy ME. 1984. Fault bars in the feathers of White-crowned sparrows: dietary deficiency or stress of captivity and handling. The Auk 101: 168 169. Lindström Å, Kivst A, Piersma T, Dekinga A, Dietz MW. 2000. Avian pectoral muscle size rapidly tracks body mass changes during flight, fasting and fuelling. Journal of Experimental Biology 203: 913 919. Littell RC, Milliken GA, Stroup WW, Wolfinger RS. 1996. SAS system for mixed models. Cary, NC: SAS Institute Inc. Machmer MM, Esselink H, Steeger C, Ydenberg RC. 1992. The occurrence of fault bars in the plumage of nestling ospreys. Ardea 80: 261 272. Mauck RA, Grubb TC. 1995. Petrel parents shunt all experimentally increased reproductive costs to their offspring. Animal Behaviour 49: 999 1008. McCullagh P, Nelder JA. 1983. Generalised linear modelling. London: Chapman & Hall. Møller AP. 1989a. Viability costs of male tail ornaments in a swallow. Nature 339: 132 135. Møller AP. 1989b. Population dynamics of a declining swallow Hirundo rustica L. population. Journal of Animal Ecology 58: 1051 1063.

FAULT BAR DISTRIBUTION AND FEATHER FUNCTION 461 Møller AP. 1994. Sexual selection and the barn swallow. Oxford: Oxford University Press. Negro JJ, Bildstein KL, Bird DM. 1994. Effects of food deprivation and handling stress on fault-bar formation in nestling American kestrels. Ardea 82: 263 267. Norberg RA. 1994. Swallow tail streamer is a mechanical device for self-deflection on tail leading edge, enhancing aerodynamics efficiency and flight manoeuvrability. Proceedings of the Royal Society of London B 257: 227 233. Pennycuick CJ. 1989. Bird flight performance. A practical calculation manual. Oxford: Oxford University Press. Riddle O. 1908. The genesis of fault bars in feathers and the cause of alternation of light and dark fundamental bars. Biological Bulletin 14: 328 371. Ritchie BW, Harrison GJ, Harrison LR. 1994. Avian medicine: principles and application. Florida: Wingers Publishing. Slagsvold T. 1982. Sex, size, and natural selection in the hooded crow Corvus corone cornix. Ornis Scandinavica 13: 165 175. Thomas ALR. 1993. On the aerodynamics of bird tails. Philosophical Transactions of the Royal Society of London B 340: 361 380. Thomas ALR. 1997. On the tail of birds. Bioscience 47: 215 225. Thomas ALR, Balmford A. 1995. How natural selection shapes bird s tails. The American Naturalist 146: 848 868. Velando A. 2002. Experimental manipulation of maternal effort produces differential effects in sons and daughters: implications for adaptive sex ratios in the blue-footed booby. Behavioural Ecology 13: 443 449. Witter MS, Cuthill IC. 1993. The ecological costs of avian fat storage. Philosophical Transactions of the Royal Society of London B 340: 73 92.