Senescence in a short-lived migratory bird: age-dependent morphology, migration, reproduction and parasitism

Authors

  • A. P. Møller,

    1. Laboratoire d’Ecologie, CNRS URA 258, Université Pierre et Marie Curie, Bât. A, 7ème étage, 7 quai St. Bernard, Case 237, F-75252 Paris Cedex 05, France, and Departamento de Ciencias Morfologicas y Biologia Celular y Animal, Universidad de Extremadura, E-06071 Badajoz, Spain
    Search for more papers by this author
  • F. DE Lope

    1. Laboratoire d’Ecologie, CNRS URA 258, Université Pierre et Marie Curie, Bât. A, 7ème étage, 7 quai St. Bernard, Case 237, F-75252 Paris Cedex 05, France, and Departamento de Ciencias Morfologicas y Biologia Celular y Animal, Universidad de Extremadura, E-06071 Badajoz, Spain
    Search for more papers by this author

A. P. Møller, Laboratoire d’Ecologie, CNRS URA 258, Université Pierre et Marie Curie, Bât. A, 7ème étage, 7 quai St Bernard, Case 237, F-75252 Paris, Cedex 05, France. Tel. (+33) 1 44 27 25 94. Fax: (+33) 1 44 27 35 16. E-mail: AMOLLER@HALL.SNV.JUSSIEU.FR

Abstract

1. Senescence reflects age-dependent changes in residual reproductive value. Annual survival rates of the barn swallow Hirundo rustica L. increased from 1- to 2-year-old individuals, but decreased among 5 years old or older individuals. Estimates of age-dependent reproductive value showed a similar pattern.

2. Longitudinal data from two long-term population studies were used to test whether a number of different measures of performance (condition-dependent morphological traits, migratory performance, reproductive success, intensity of parasitism) changed among individuals when reaching old age.

3. The length of the outermost tail feathers (a secondary sexual character) decreased among old individuals, while two measures of individual developmental instability increased with age. Migratory performance decreased in old barn swallows as reflected by a delay in spring arrival at the breeding grounds. Reproductive performance measured as seasonal reproductive success decreased with age. The intensity of infestations with an haematophagous mite and a mallophagous ectoparasite increased among old barn swallows.

4. These results suggest that the condition-dependent secondary sexual character, developmental stability, and measures of migratory and reproductive performance deteriorated, and the frequency of parasitism increased among old individuals. Ageing was thus associated with a general deterioration of performance.

Introduction

Senescence is the decrease in age-dependent residual reproductive value in species with separate germ line and soma (Fisher 1930), usually giving rise to a deterioration of the phenotype with an associated increase in mortality rate. Such apparently maladaptive increase in mortality among experienced individuals, even though they have been exposed repeatedly to episodes of intense natural and sexual selection, originally posed a problem to the theory of natural selection. Deterioration of the phenotype with increasing age was subsequently assumed to be a sign of senescence and a number of different theories was proposed to account for its evolution (Rose 1991; Partridge & Barton 1993). Medawar (1952) suggested that deleterious mutations may accumulate during the lifespan of an individual and cause an increase in age-dependent mortality, in particular among long-lived species with a large number of individuals reaching old age. A second major hypothesis suggested that mutations with beneficial effects early in life, such as an increased rate of reproduction, but with an associated later cost, would not be selected against because the subsequent mortality would not, or only to a small extent, affect the fitness of an individual (Fisher 1930; Williams 1957). Age-specific selection pressures were later shown to be due to age-specific sensitivity of fitness to changes in reproductive or survival rates (Hamilton 1966). Such pleiotropic genes with beneficial effects early in life and detrimental effects later on should generally be associated with a deterioration of overall performance at old age, and this process could be accelerated by early costly activities such as a high rate of reproduction. A third hypothesis states that deleterious mutations expressed only late in life are able to accumulate because of their negligible effects on overall fitness, reflecting the low probability of survival until the age at which they are expressed (Rose 1991; Partridge & Barton 1993).

Previous studies of senescence in free-living organisms are almost entirely restricted to demonstration of age-dependent mortality, reproduction and reproductive value (e.g. Newton & Rothery 1997). Although such studies are necessary for demonstration of the existence of senescence, they do not provide any clues to the mechanisms of senescence. These are still a subject of intense scientific investigation. The exact mechanisms may involve an inability to repair damaged strands of DNA, a deterioration of the immune system and, hence, an increase in the frequency of parasitism and the negative effects of mutations at late stages of the life cycle.

Birds are particularly appropriate study organisms for research on senescence because birds have very high survival rates for a given body size compared to mammals (Williams 1957; Comfort 1979; Nesse 1988; Promislow 1991). This result is surprising given the very high metabolic rates of birds, and, thus, the potentially high rates of damage to DNA, as compared to mammals. Relatively many individuals of birds therefore reach old age when the effects of senescence are expected to be particularly detectable. This fact has not been exploited to any great extent since insects and mammals in particular have been the target of most studies of senescence (Rose 1991).

A number of different measures of performance are suitable yardsticks against which a deterioration of performance at old age can be compared.

1. Sexual selection arises from competition among individuals for access to mates and ultimately the fertilization of their eggs, and it has led to the evolution and the maintenance of extravagant secondary sexual characters (Darwin 1871). The time and energy costs involved in the production and the maintenance of secondary sexual characters, as well as their great intricacy of design, render them particularly susceptible to condition-dependent expression (Andersson 1994). Any deterioration in the ability to acquire resources necessary for the production and the maintenance of secondary sexual characters, as well as any mutation that affects their expression, will invariably result in the production of a phenotype of poor quality. Hence, we would predict that secondary sexual characters should decrease in size at old age for species that re-grow their secondary sexual characters annually or growth increments should decrease for species that maintain a single secondary sexual character throughout life.

2. Developmental stability reflects the ability of individuals to undergo stable development of their phenotype under given environmental conditions (Møller & Swaddle 1997). Deviations from this condition of stability are caused by a wide range of environmental (for example, nutrition, parasitism, temperature and population density) and genetic factors (for example, mutation, inbreeding and hybridization; review in Møller & Swaddle 1997). Such deviations are traditionally measured in terms of degree of randomly directed deviations from perfect bilateral symmetry, so-called fluctuating asymmetry, or the frequency of major phenotypic deviants, so-called phenodeviants (Møller & Swaddle 1997). Developmental performance measured as fluctuating asymmetry is directly associated with performance in intra- and interspecific interactions like competition, predation and parasitism (Møller & Swaddle 1997). A number of different fitness components, such as growth rate, survival, fecundity and mating success have been shown to be consistently higher in more symmetric individuals in a diverse array of species ranging from plants to animals (reviewed in Møller & Swaddle 1997). Although most organisms only develop their phenotype once during early parts of life, birds are very suitable organisms for the study of repeated developmental performance because feathers are regrown during the annual moult. Hence, the developmental performance of the same genotype can be studied when the potential effects of senescence start to influence the ability to control ontogenetic processes. Since tail length is a condition-dependent secondary sexual character with expression depending on foraging conditions in the winter quarters, parasite loads and radiation (Møller 1990, 1991, 1993), and since fluctuating asymmetry depends on the condition of individuals as determined by foraging conditions in the winter quarters, parasite loads and radiation (Møller 1992, 1993 and unpublished data), we predict that asymmetry should be inversely related to tail length during the lifespan of an individual.

3. Long distance migration is a common phenomenon among birds, with many species flying thousands of kilometers twice per year (Alerstam 1982; Berthold 1990). Migration involves both the expenditure of large amounts of time and energy, and the ability to find appropriate places for staging, wintering and reproduction. Not surprisingly, the ability of individual birds to return to their breeding grounds is closely associated with indicators of high quality such as the expression of secondary sexual characters (Møller 1994a). Both the amount and the direction of migration are under genetic control (Berthold 1990), and migratory performance can obviously be directly affected by a general deterioration of the phenotype, as well as the accumulation of deleterious mutations. The prediction is therefore that arrival is delayed among old individuals.

4. Senescence reflects the decrease in residual reproductive value with age and is closely associated with life-history traits such as age-specific survival rates and various measures of reproduction (Rose 1991). There is increasing evidence that life history traits such as timing of reproduction and clutch size are the result of individual optimization with respect to condition (Högstedt 1980; Tinbergen & Daan 1990; Brinkhof 1995; Verhulst 1995). Any age-specific deterioration of condition due to senescence should result in a decrease in reproductive performance.

5. Parasitism and genetically based diseases have long since been implicated in the expression of senescence (Medawar 1952; Williams 1957). The reasons are that a typical sign of senescence is an increased susceptibility to parasitism and the expression of genetically-based diseases, such as cancers (review in Miller 1996). Immune function has several times been shown to deteriorate among old individuals, in particular changes in T lymphocyte populations due to the replacement of virgin T cells by memory T cells, leading to reduced abilities to fight off parasites and prevent the expression of disease (review in Miller 1996). Thus, individuals should, when old, have a larger intensity of parasites than when young.

The main aims of this study were to investigate to which extent senescence occurred in a short-lived migratory bird, the barn swallow Hirundo rustica, using a number of different measures of performance. More specifically, we tested the basic assumption arising from the definition of senescence that age-specific mortality increased and reproductive value decreased among old individuals. Secondly, we tested whether a number of different measures of performance deteriorated among old individuals: a decrease in the size of a secondary sexual character, an increase in developmental instability as estimated from fluctuating asymmetry, and a decrease in migratory and reproductive performance, and parasite resistance as reflected by infestations with two species of ectoparasites. This study is the first to investigate in a free-living organism a wide range of measures of deteriorating performance other than reproduction among individuals reaching old age. Therefore, the study provides information on the potential mechanisms that are involved in generating age-specific mortality and reproductive values.

The barn swallow is a ≈20 g migratory, insectivorous bird that breeds in large parts of temperate and subtropical zones of the northern hemisphere. Males and females have similar morphology with the exception of the outermost tail feathers which are longer in males due to a directional female mate preference (review in Møller 1994b). The mating system is social monogamy with frequent cases of extra-pair paternity, mainly by females mated to short-tailed males (Saino et al. 1997). Migration, reproduction and ability to resist parasites are all associated with long tails in male barn swallows (Møller 1994b). The barn swallow is therefore a very suitable study organism for investigations of senescence because a large number of measures of performance is directly associated with the expression of a single phenotypic trait.

Methods

We studied barn swallows at two sites (Kraghede, Denmark, and Badajoz, Spain) during 1982–96 and 1990–96, respectively. These two populations have been followed carefully in detailed population studies with more than 98% of all adult birds being captured annually. Due to the very low degree of breeding dispersal of barn swallows, with less than 1% of all adults ever changing breeding sites within or between seasons (a conclusion based on more than 1000 adults recaptured), survival rate can readily be estimated from the disappearance of adults from 1 year to the next (Møller 1994b; A. P. Møller & F. de Lope unpublished data). Less than 2% of all birds were recaptured in 1 year, but missed in the previous year (Møller 1994b; A. P. Møller and F. de Lope unpublished data). Adult barn swallows were assigned an age of 1 year when first recorded as breeding. This assumption was supported by more than 200 local recruits all being captured as breeders in their first year of life, and never at an older age when first captured (Møller 1994b; A. P. Møller and F. de Lope unpublished data). Adults were sexed based on the shape of their cloacal protuberance (Svensson 1984) and their subsequent reproductive behaviour.

Adults were captured in mist nets upon spring arrival and a number of morphological measures were recorded annually: the length of the right and left flattened wing, right and left outermost tail feather, and central tail feather to the nearest mm with a ruler, the length of the right tarsus to the nearest 0·01 mm with a digital caliper, and body mass with a Pesola spring balance to the nearest 0·1 g. Morphological measures were highly repeatable as determined from repeated measurements of the same individuals in the same breeding season (Møller 1991, 1994b; F. de Lope & A. P. Møller unpublished data). The morphological measurements were used to calculate absolute individual fluctuating asymmetry of wing length and outermost tail length by simply using the unsigned difference between the length of the right and the left character values. Signed left-minus-right wing and tail lengths were normally distributed with a mean value not deviating significantly from zero (see also Møller 1994c).

A measure of migratory performance was obtained in the Danish population by recording the first day when an individual was present as the date of arrival. Since breeding sites were inspected daily for arrival of adult barn swallows during spring, this provides a reliable estimate of the ability of early migration and arrival at the breeding grounds. This measure of individual migratory performance is significantly repeatable among years (Møller 1994a).

Reproduction was monitored by regular visits at least once per week during the breeding season, and more frequently during laying and hatching. Brood size was estimated as the number of nestlings produced in each of the up to three clutches. Seasonal reproductive success was recorded as the total number of fledglings produced in the up to three different annual clutches.

Parasites were recorded in adult barn swallows and their nests. The number of mallophaga of the species Hirundoecus malleus L. was recorded from the tail feathers of adults when first captured each year. Parasite abundance was estimated as the number of holes in the feathers chewed by this feather parasite. This is a reliable estimate of the relative intensity of infestation (Møller 1990). The intensity of infestation with the haematophagous mite Ornithonyssus bursa Berlese in the nest was determined on the first day after fledging of nestlings by placing a hand on the rim of the nest for 10 s and then recording the number of mites in orders of magnitude as 0, 10, 100, 1000 or 10 000. This method provides a reliable measure of the intensity of infestation of the nest as determined from extraction of mites from a sample of nests (Møller 1990), and also of the infestation of adult breeders associated with the nest (Møller 1990).

We included adult barn swallows in this study if the birds had reached at least 5 years of age. This leaves a relatively small cohort of individuals with measures of performance unaffected by selection since each year class contained exactly the same individuals. If birds were older than 5 years, we simply calculated the mean performance for five years and older of that individual. Annual survival rate was calculated as the proportion of adults reaching a specific age divided by the number of individuals in the previous age class. The standard error of this estimate was calculated assuming that survival rate was a variable with a binomial distribution (Sokal & Rohlf 1995).

We calculated reproductive value of an individual at age a according to Stearns (1992) as

image

summed over all age classes from a to A, where A is the maximum age, lx is the survival to age x and mx is the mean number of fledglings raised by individuals of age x. The standard error of this estimate was calculated according to the equation in the appendix in Newton & Rothery (1997).

The data set consisted of various measures of annual performance of the same individuals throughout their lives. The measures of age-specific performance of adult barn swallows were therefore compared using repeated measures analysis of variance. We tested whether the pattern of age-specific performance differed between the sexes by inclusion of the interaction terms sex × age and sex × age2 in the analyses of variance. These interaction terms were not statistically significant in any of the eight models (sex × age: 0·17 < P < 0·88; sex × age2: 0·11 < P < 0·99), although the power of these tests is low because of the relatively small sample sizes. Hence, we have not considered sex effects in the analyses reported in the Results section. The abundance of mallophaga was log10 (x + 1) transformed before calculations. We performed post-hoc tests by assuming that performance in age classes 1 and Ð5 would be similar, but would differ from performance in age classes 2–4. The rationale for this procedure is that young individuals generally perform less well than older individuals because of inferior experience, for example, in terms of foraging and this may affect parasite loads, migration and reproduction (Williams 1957; Alerstam 1982; Berthold 1990; Stearns 1992). Old individuals were predicted to perform poorly because of the effects of senescence. Hence, we used the simplest possible post-hoc test by assuming that yearling and old individuals would perform equally poorly, while individuals of intermediate age classes would perform better. All statistical tests were two-tailed and the level of significance 5%.

Results

Annual survival rate of barn swallows varied significantly with age, with pairwise comparisons differing at the 5% level (Fig. 1a; t-tests). There was an initial increase in survival from year 1 to year 2, followed by a stable level among 2–4-year-old individuals and a subsequent decrease among individuals that were five years or older. Reproductive value showed a similar pattern with a maximum of 11·63 at two years of age and a subsequent decrease to 3·00 at 6 years (Fig. 1b). A polynomial regression based on the mean values for each age class revealed a highly significant relationship (F = 72·19, d.f. = 2,4, r2 = 0·96, P = 0·0007), with both the linear and the quadratic terms being statistically significant [linear coefficient: 1·99 (SE = 0·53), P = 0·0088; quadratic coefficient: −0·51 (SE = 0·07), P = 0·0016]. Hence, the statistical analysis supported the pattern in Fig. 1.

Figure 1.

Figure 1.

Age-dependent (a) annual survival rates (+SE) and (b) reproductive values (+SE) of Danish barn swallows. Numbers are number of individuals reaching each age class. SEs are too small to be visible for survival of age classes 1 and 2.Ücc35p

Figure 1.

Figure 1.

Age-dependent (a) annual survival rates (+SE) and (b) reproductive values (+SE) of Danish barn swallows. Numbers are number of individuals reaching each age class. SEs are too small to be visible for survival of age classes 1 and 2.Ücc35p

We recorded a total of 53 individuals (26 males and 27 females) that became 5 years old or older during our population studies. The length of the outermost tail feathers varied significantly among age classes (Table 1). There was an initial significant increase in tail length from year 1 to year 2, and from year 2 to year 3, with a non-significant decrease among old individuals (Fig. 2a).

Table 1.  Repeated-measures analysis of variance of morphology, migration, reproduction and parasitism of barn swallows in relation to age
Variance componentMSFd.f.PP*
  1. P *: probability from post-hoc test of the comparison between means based on the hypothesis that performance in age classes 1 and Ð5 is similar, but differs from performance in age classes 2–4.

Tail length
Among individuals613·4456·60520·0001 
Within individuals
Age147·7818·0140·00010·0001
Residual8·20 208  
Tail asymmetry
Among individuals9·022·43520·0001 
Within individuals
Age21·906·5140·00010·01
Residual3·36 208  
Wing length
Among individuals48·7536·14520·0001 
Within individuals
Age10·128·5840·00010·01
Residual1·18 208  
Wing asymmetry
Among individuals0·591·80520·002 
Within individuals
Age1·073·4340·00960·03
Residual0·31 208  
Arrival date
Among individuals0·732·32250·0017 
Within individuals
Age0·772·5940·0410·02
Residual0·30 100  
Reproductive success
Among individuals7·631·08520·35 
Within individuals
Age75·2413·0140·00010·0001
Residual5·78 208  
Intensity of mite infestation
Among individuals1·051·72520·031 
Within individuals
Age2·073·7340·00710·01
Residual0·55 208  
Intensity of mallophaga infestation
Among individuals0·573·00520·0001 
Within individuals
Age1·197·0040·00010·001
Residual0·17 208  
Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Figure 2.

Figure 2.

Morphology, migration, reproduction and parasitism of barn swallows in relation to age (years). Values are means (+SE). For arrival date 1 = 1st May. Sample size was 53, with the exception of arrival date for which it was 26.Üjf

Although wing length was much less variable than tail length from one year to the next, the age-dependent pattern of variation was similar to that of the secondary sexual character (Table 1). Wing length initially increased significantly from year 1 to year 2, with a significant reduction among individuals of the oldest age class (Fig. 2c).

Fluctuating asymmetry in tail and wing length varied significantly among age classes (Table 1). Tail length asymmetry demonstrated an initial significant decrease from year 1 to year 2 followed by a significant increase among birds that were 5 years or older for both wing and tail asymmetry (Fig. 2b,d).

Migratory performance measured as the date of arrival varied significantly among age classes (Table 1). Spring arrival was later among barn swallows that were 1 year or 5 years and older than among individuals of intermediate age (Fig. 2e).

Reproductive success of the barn swallow varied significantly among age classes (Table 1). There was an initial increase in reproductive success between age class 1–2 years and 3 years with significant subsequent decreases from year 3 to year 4, and 5 years old or older (Fig. 2f).

The intensity of infestation with the two ectoparasites varied significantly among age classes of barn swallow hosts (Table 1). Barn swallow nests of 1–2-year-old birds and individuals that were at least 5 years old held more mites than barn swallow nests with owners of intermediate ages (Fig. 2g). Similarly, the abundance of mallophaga in the feathers of adult barn swallows decreased from the first and second to the third year of the life of the host with a subsequent increase among 4- and 5-year-old hosts (Fig. 2h).

Discussion

Senescence is the age-dependent decrease in residual reproductive value among individuals reaching old age. It has been hypothesized to arise as a consequence of ageing due to the accumulation of deleterious mutation and/or the negative pleiotropic effects late in life of alleles with a beneficial effect during early life stages (Medawar 1952; Williams 1957; Hamilton 1966; review in Rose 1991). Although senescence has been recorded in a large number of studies of animals in captivity, there is little available information on the relative importance of senescence in the wild, let alone the mechanisms giving rise to an increase in age-dependent mortality and a reduction in residual reproductive value. The phenomenon of ageing has recently attracted considerable attention from evolutionary biologists, and an increasing number of studies of free-living organisms have demonstrated effects of senescence (Partridge & Barton 1993). Danish barn swallows experienced an increase in mortality rate among individuals that were at least 5 years old (Fig. 1a), and since reproductive success also decreased among birds of this age class (Fig. 2f), residual reproductive value decreased disproportionately among old birds (Fig. 1b). Hence, there is little doubt that barn swallows experienced senescence, although the very small proportion of adults reaching an age of 5 years certainly suggests that the proportion of senescent individuals is very low. This is in accordance with the frequency of individuals experiencing effects of senescence being disproportionately large in species with low mortality rates and small in species with high mortality rates such as the barn swallow.

The mechanisms and consequences of senescence have received considerably less attention than the mere documentation of its presence and extent. The present study differs from the entire literature on senescence in free-living organisms by providing a number of potential mechanisms that may give rise to an age-specific deterioration in performance. Secondary sexual characters have been found often to show condition-dependent expression, as demonstrated by many studies showing positive associations between presumed markers of good condition, such as size and age, and the size of secondary sexual characters, and experimental demonstrations of reduced condition giving rise to less extravagant secondary sexual characters (reviews in Andersson 1994; Johnstone 1994). Previous experimental studies of the barn swallow are consistent with this literature by demonstrating a reduction in the size of a secondary sexual character (tail length) the subsequent year after experimental elongation (Møller 1989). The reduction in tail length among old barn swallows, as shown in the present study (Fig. 2a), is consistent with this experiment by suggesting that tail length is reduced when birds reach old age. We are only aware of one other study demonstrating a decrease in the size of a secondary sexual character among individuals reaching old age (red deer Cervus elaphus: Clutton-Brock et al. 1982).

Fluctuating asymmetry is a measure of developmental instability reflecting the adverse effects of genetic and environmental factors on developmental processes (review in Møller & Swaddle 1997). If tail and wing length are condition-dependent traits, as suggested by the review in the Introduction, we should expect an inverse relationship between asymmetry and size of morphological characters. Hence, fluctuating asymmetry should demonstrate a U-shaped relationship with age. Mean fluctuating asymmetry in wing and tail length was high among yearlings, but decreased subsequently. This initial decrease is likely to reflect an improvement in condition. The increase in fluctuating asymmetry in tail and wing length among barn swallows reaching old age (Fig. 2b,d) must be due to deterioration of body condition. A previous study of the barn swallow has shown that experimental elongation of tail feathers resulted in increased asymmetry in the subsequent year (Møller 1992), implying that asymmetry arises as a consequence of deteriorating condition. Although the changes in fluctuating asymmetry are small, they may have severe consequences for locomotion and manoeuvrability. For example, Swaddle (1997) has shown that even minute changes in wing asymmetry following moult dramatically affect measures of flight performance in starlings Sturnus vulgaris L. Age-specific changes in asymmetry may thus affect flight performance and thereby arrival date in this long-distance migrant. Furthermore, flight performance may directly affect foraging efficiency and, hence, body condition. This may have consequences for reproductive success which is known to be condition-dependent in many species of birds (Högstedt 1980; Tinbergen & Daan 1990; Brinkhof 1995; Verhulst 1995). Even parasite load may be a direct consequence of decreased foraging efficiency due to increased morphological asymmetry because foraging competes with time spent preening, and because several aspects of immune function are condition-dependent (Chandra & Newberne 1977; Gershwin, Beach & Hurley 1985).

Long-distance migration is one of the most time and energy consuming activities of free-living organisms. Many bird species travel several thousand kilometers twice a year between the breeding sites and winter quarters, and any deterioration of the phenotype will invariably result in poor migratory performance. Barn swallows arrived earlier with increasing age, as shown in many other birds (Alerstam 1982; Berthold 1990), but old individuals arrived later than young ones (Fig. 2e). This finding indicates that performance decreased among individuals of very old age.

Energy expenditure during reproduction reaches a very high level, and therefore it is perhaps not surprising that reproductive performance of barn swallows decreased among individuals of old age (Fig. 2f). The mechanism giving rise to such a decrease in reproductive performance may be linked to several of the phenotypic traits that changed with age. For example, late arrival generally results in delayed start of reproduction and a reduced seasonal reproductive success due to a reduced probability of rearing multiple broods during the same season (Møller 1994b). Increased asymmetry in wings and tails may directly interfere with flight and thus affect the accumulation of resources necessary for reproduction and successful rearing of young (Swaddle 1997). The increased levels of parasitism among old individuals may also directly affect the reproductive success of hosts, since the haematophagous mite Ornithonyssus bursa has been demonstrated experimentally to reduce the quantity and the quality of offspring produced (Møller 1990).

Parasitism and disease resistance has been linked to senescence for a long time, mainly because old individuals have been found to suffer from diseases more often than younger ones. Although this mechanistic explanation for senescence has not been favoured recently, there is a considerable literature on humans and domesticated mammals which suggests that senescence is associated with a deterioration of various aspects of immune function (review in Miller 1996). We found increased intensities of infestation among individual barn swallows in their first year and when 5 years old or older (Fig. 2g,h). This increase in infestation among old individuals is consistent with a reduced ability of hosts to defend themselves against parasites when reaching old age. Whether this effect was due to differences in exposure, behavioural defence or immune defence remains unknown.

The relative importance of mutation accumulation and pleiotropic effects in generating the effects of senescence shown in this study is not readily disentangled. However, the question is not untractable since mutation rates of microsatellite markers in the barn swallow are so high (Primmer et al. 1996) that the rate of accumulation of mutations over the lifespan of individuals can be directly estimated.

In conclusion, a number of different measures of performance in the barn swallow deteriorated among very old individuals and were directly related to an increase in age-specific mortality rate.

Acknowledgements

G. Sorci kindly provided constructive comments. The research was supported by grants from the Danish Natural Science Research Council to A. P. M. and from DGICYT PB95–0020 to F. de L.

Received 14 May 1997;revisionreceived 5 May 1998

Ancillary