The effects of territory quality on age-dependent reproductive performance in the northern wheatear, Oenanthe oenanthe

ANIMAL BEHAVIOUR, 2001, 62, 379–388 doi:10.1006/anbe.2001.1754, available online at http://www.idealibrary.com on

The effects of territory quality on age-dependent reproductive performance in the northern wheatear, Oenanthe oenanthe TOMAS PA } RT

Department of Conservation Biology, The Swedish University for Agricultural Sciences (Received 3 August 2000; initial acceptance 27 November 2000; final acceptance 30 January 2001; MS. number: A8845)

Age differences in access to high-quality resources and its relation to nest predation risk have generally been neglected in efforts to explain why old individuals have a higher reproductive performance than young ones. I used long-term population data to investigate age- and habitat-specific nest predation risk and reproductive success of male northern wheatears breeding in a heterogeneous agricultural landscape. Old (d2 years) male wheatears had higher reproductive success than yearling males because their nests were less likely to be predated, they produced more fledglings from successful broods and they were more likely to start a new breeding attempt following a complete failure. Also, old males arrived and bred earlier than yearlings. Old wheatears were more likely to breed in habitats with a permanently short field layer, while yearlings mainly bred in habitats with a field layer gaining in height during the incubation– nestling period. Nest predation risk was significantly lower and fledgling production among successful breeders was higher in territories with a permanently short as compared to growing field layer. Data on order of habitat occupation suggested that yearlings were not as good as older males at identifying territories with a permanently short field layer (as reflected by a significant interaction between male age and order of occupancy). Both cross-sectional and longitudinal data suggest that an age difference in territory field layer height was a major cause of age differences in reproductive performance in wheatears. Other factors, such as differential survival, breeding time, local familiarity and increased foraging and breeding skills either had small effects or did not significantly explain the observed age differences in reproductive performance. I conclude that nest predation and an age-related access to high-quality habitats are likely to be of great importance for the observed age differences in reproductive performance in the northern wheatear and possibly many other bird species. 

competence early in life (Forslund & Pa¨rt 1995). Because of the problems of measuring reproductive effort, no study has yet been able to show unambiguously that reproductive effort increases with age early in life. However, several studies of birds suggest that enhanced competence may be especially important (Newton 1989; Forslund & Pa¨rt 1995), and there seems to be a general agreement that these improvements mainly are caused by increased breeding experience, mate experience and foraging ability (e.g. Newton 1989; Saether 1990; Fowler 1995). Several studies of breeding habitat selection of birds have suggested that old individuals (second-time breeders or older) often breed in better habitats than first-time breeders (Cody 1985; Bernstein et al. 1991; Newton 1992). Furthermore, a number of dispersal studies have indicated that young birds are often relegated to poorer nesting sites, and only acquire better sites as they grow older (Greenwood & Harvey 1982). Thus, an improved reproductive performance with age may also be

Most animals that reproduce on several occasions in their lifetime display age-specific patterns of reproductive performance, where an increase in reproductive performance early in life is most pronounced between the first two reproductive events (see Clutton-Brock 1988; Newton 1989; Saether 1990; Charlesworth 1994). Three main groups of factors causing this initial increase in reproduction have been identified, namely (1) progressive appearance and disappearance of phenotypes (e.g. differential survival), (2) age-related improvements in competence and (3) optimization of reproductive effort (for a review, see Forslund & Pa¨rt 1995). Most studies of birds suggest differential survival to be of minor importance for the observed patterns of age-related reproductive performance (Forslund & Pa¨rt 1995). Increased reproductive effort with increasing age during the first years of life either require early senescence or age-related improvements in Correspondence: T. Pa¨rt, Department of Conservation Biology, SLU, Box 7002, SE-750 07 Uppsala, Sweden (email: Tomas.Part@ nvb.slu.se). 0003–3472/01/080379+10 $35.00/0

2001 The Association for the Study of Animal Behaviour

379



2001 The Association for the Study of Animal Behaviour

380

ANIMAL BEHAVIOUR, 62, 2

influenced by priority access to high-quality habitats with increasing age. Nest predation risk may be an important component of habitat quality, since it may be related to habitat properties (e.g. Martin 1988, 1992) and often is the most important single causal factor explaining complete breeding failures in birds (Ricklefs 1969; Martin 1992). Also, nest predation risk may be related to differences in individual behaviour affected by skill and experience (e.g. Martin 1992; Haskell 1994). Consequently, one may expect nest predation risk to be associated with the age of breeding individuals. Few studies, however, have investigated the potential impact of habitat quality and nest predation risk on age-specific reproductive patterns, probably because many studies have been made in habitats with a naturally low or artificially reduced risk of nest predation (i.e. nestbox studies; Møller 1989). Still, unequal distribution of age classes in a heterogeneous environment and nest predation itself can potentially influence many life history traits (Martin 1992; Sutherland 1996), and may therefore be important factors shaping age differences in reproductive performance. By using long-term cross-sectional and longitudinal data from a population of the migratory northern wheatear, I investigate potential causes for age-specific reproductive performance and its relation to nest predation risk and territory habitat structure. Also, I test whether differential survival of phenotypes, habitat selection patterns, breeding time, local familiarity and other individual improvements in competence (e.g. foraging skills and breeding experience) may explain the observed age differences in reproductive performance. METHODS

Study Species The northern wheatear is a small (21 g), migratory insectivorous bird species that winters in Africa south of Sahara and breeds in Europe, Asia and the northern parts of North America. In my study area near Uppsala, Sweden, males arrive at the breeding grounds on average a few days before females; that is from mid-April to mid-May. Older birds (d2 years) arrive on average a week before yearlings. Wheatears usually nest in ground cavities, often under stones, in stone piles and in stone walls, but about 20% of the pairs in this study area nested under roof tiles (1–6 m above ground). Egg laying starts in early May, incubation lasts for 13 days, and nestlings stay in the nest for about 15 days. Most food is taken from the ground by a hopping technique or by flights from low perches. The height of the field layer is probably the most important vegetation determinant of habitat quality for breeding wheatears, since a short field layer increases prey availability, decreases foraging costs and thus increases reproductive success (Conder 1989; Tye 1992).

Study Population This study was conducted between 1993 and 1999 in a 60-km2 large study area situated in farmland southeast of

Figure 1. Schematic picture of the tube trap used to capture adult northern wheatears at nest cavities. The trap had a one-way swing entrance and was closed with mist netting at the opposite end. The shaded rectangle is a small plastic tube through which a U-shaped wire was inserted, allowing the wire to move freely in only one direction.

Uppsala, Sweden (5950 N, 1750 E). About 120–180 pairs of wheatears breed annually in the study area. Since wheatears prefer to forage in habitats with a short field layer, most pairs breed in grazed or mowed grasslands (about 70%), although some pairs breed in crop fields (16%), farmyards (13%) and abandoned grasslands (1%). From mid-April to early July I visited all potential breeding sites at least every fifth day to record arrival date and location of established pairs. At the time of nest building and hatching, we (my field assistants or I) visited each pair at least every third day. When a pair was seen feeding young, we located the nest and determined hatching date by backdating from nestling age. We estimated age of nestlings by comparing photographs of known-aged nestlings (1–10 days old, N=10 nests with known hatching date in 1992). Because most nests were found within 3 days after hatching, uncertainties due to differences in growth rate caused by territory quality or weather were minimal. We also found some nests during nest building, egg laying and incubation. I determined clutch size in nests checked at incubation or at hatching (i.e. within 1 day of hatching of the first egg). I calculated brood reduction, as estimated by fledging success (i.e. number fledged/hatched young), only for nests with known brood size at hatching. We marked nestlings with colour leg rings and an aluminium leg ring (in a unique combination) when they were at least 4 days old. We determined the number of fledged young by counting fledged young and by checking the nests for dead young after the time of fledging. Small dead nestlings were carried away by the parents, but nestlings older than 3–4 days were left in the nest cup. When all nestlings died before the age of 4 days, usually the last nestling to die was left in the nest cup. We captured breeding adults mainly in the central part of the study area (about 60 pairs/year) using mist nets, or using a tube trap placed at the entrance to the nest cavity. The tube trap (length 150 mm, diameter 50 mm) had a one-way swing entrance (i.e. two thin wires) and was closed at the opposite end with mist netting (see Fig. 1). We marked all adults with a unique combination of colour leg rings and an aluminium leg ring, and took morphometrical measures at the time of capture. We aged all established males in the field as yearlings (brown wings) or older (black wings; see Svensson 1992), and my data on males of known age (i.e. ringed as nestlings) showed no exceptions to this rule. Only a small fraction of the unringed breeding females were aged as yearlings or older in the field. However,

PA } RT: AGE, NEST PREDATION AND TERRITORY QUALITY

about 95% of the females caught could be classified as either yearlings (worn light brown remiges and coverts, worn tail feathers, pale inside of upper mandible) or older (d2 years; darker, less worn feathers; see Svensson 1992; Jenni & Winkler 1994). There was a strong positive association between male and female age within pairs; old males were mated mainly with old females (86% of 385 old males), while yearling males were mated mainly with yearling females (73% of 263 yearling males;
21 =246.8, P<0.0001). We recorded breeding success (i.e. failed or successful) for all established pairs but obtained fledgling production (i.e. number fledged young in successful nests) only from 72% of the successful breeding attempts, because some nests were impossible to reach. I estimated survival rates only for males breeding in the central parts of the study area (i.e. about 6 km from the periphery of the whole study area). I considered males to have died when they were not observed in the 2 following years. This estimate of survival is very reliable because dispersal of adults was well within the study area (median distance=0 m; range 0–4500 m) and no breeding male escaped detection in any year when he survived more than 2 years after his first year of reproduction. Many breeding attempts completely failed, and of 146 nests that were inspected following a failure, 85% were depredated; the others failed because of starvation or because of death of at least one parent). Because not all nests that failed were found, I used complete failures as an estimate of nest predation risk. Most depredations were probably caused by stoat (Mustela erminea), weasel (Mustela nivalis) and adder (Vipera berus), all of which occur throughout the study area and which have been observed depredating nests (unpublished data). Each year about 5–10 nests were also depredated by badgers (Meles meles) and red fox (Vulpes vulpes). However, other nest predators such as corvids (Corvidae) were not able to reach most of the nests.

Ethical note Breeding adults were caught during midday hours under good weather conditions to reduce potential effects of lost feeding of nestlings. Tube traps and mist nets were checked at least every 15 min and in no case did a caught bird die. After release, adults resumed feeding their nestlings within 30 min. The maximum time a trap was set in a nest was 45 min, but in most cases birds were trapped within 15 min. Catching adults did not affect nestling survival to fledging (i.e. fledging success; analysis of covariance (ANCOVA): year*field layer height*male age*hatching date*catching: catching effect: F1,362 =0.26, P=0.60). Nestlings were ringed only under good weather conditions. Nestlings older than 9 days may leave the nest when ringed. I therefore mainly ringed small nestlings (i.e. 4–8 days old). Because nestling tarsi grow very fast, ordinary sized rings could be used on nestlings as young as 4 days old. Although the tarsi continue to grow up to an age of 9 days, I observed no case when rings had caught on objects or caused injuries. However, two out of 2485 ringed nestlings died because of accidents.

Height of Field Layer in Territory Wheatears mainly forage by running and jumping in vegetation shorter than 5 cm and the average territory size of wheatears is about 2 ha (Conder 1989). Based on this information and based on the fact that field layer height seems to be an important determinant of territory quality of wheatears (Tye 1992), I grouped territories (i.e. a 2-ha circle around the nest site) into two distinct classes. (1) Short: the field layer was permanently shorter than 5 cm during the breeding period. Grazed pastures and some farmyards with cut lawns were included in this category. (2) Tall: the field layer grew to 15 cm or more during late incubation and nestling care. Pastures not grazed in spring, leys (cultivated grasslands), crop fields, and some farmyards, old abandoned pastures and old fallow fields were included in this category. In most cases (99%), however, the field layer was shorter than 5 cm at the time of territory establishment (i.e. late April– early May). The few nest sites that were located at the border between the two habitat types were omitted from the analyses.

Statistics I treated each breeding attempt as an observation in cross-sectional comparisons. Thus some males may occur repeatedly in these analyses. However, in the subset of ringed birds, male reproductive success was uncorrelated between years (spearman rank correlation: rS =
0.005, N=111 males, P=0.96), partly because males experienced new breeding conditions between years (e.g. new mates, territories and breeding dates). I used logistic regressions to analyse nominal data (e.g. probability of nest predation and second breeding attempts) and the reported chi-square values refer to likelihood-ratio chi-square tests. I used analysis of variance (ANOVA) or covariance (ANCOVA) models to analyse clutch size, fledging success (i.e. arcsine of number fledged/number hatched) and number of fledged young in successful breeding attempts. For multivariate models, I started with a model including all interaction terms. I then deleted all nonsignificant interaction terms (P>0.05), starting with those of highest order. I analysed longitudinal data on within-individual changes to assess differences in reproductive performance of males between 1 and 2 years of age. Within-individual differences in breeding success (i.e. 0 or 1) could obtain values of
1, 0 and +1, and hence were analysed with ordinal logistic regression. All analyses were performed on JMP (version 3.2; SAS Institute 1999) on balanced data sets. Sample sizes vary among tests because of missing values.

RESULTS Yearling and older male northern wheatears differed markedly in arrival time, reproductive performance and habitat choice. Older males arrived earlier, more often chose territories with a permanent short field layer,

381

382

ANIMAL BEHAVIOUR, 62, 2

Table 1. Comparison of habitat choice and reproductive performance of yearling and old (≥2 years) male northern wheatears Yearling

All pairs: Date of arrival† %Nest predation %Second attempts‡ %In short field layer§ Successful pairs: Date of hatching† Clutch size Number fledged Fledging success**

Old

N

Mean

SE

N

Mean

SE

P*

259 391 123 394

7.36 33.5 7.3 47.7

0.50

416 586 143 588

0.09 25.7 23.8 62.2

0.39

<0.0001 0.005 <0.0001 <0.0001

246 123 191 121

38.40 6.00 4.62 0.88

0.36 0.07 0.10 0.02

403 189 310 188

34.98 6.17 5.14 0.93

0.27 0.05 0.08 0.01

<0.0001 0.04 <0.0001 0.014

*P values refer to ANOVAs (year, male age) for all variables except for nominal ones where logistic regression was used. All models were significant (P<0.01). †May 1=1. ‡Percentage of pairs making a second breeding attempt following a complete failure. §Percentage of pairs breeding in territories with a permanently short field layer (as opposed to a field layer growing tall). **Number fledged/hatched young.

started to breed earlier, experienced a lower nest predation risk, fledged more young in successful attempts and more often attempted re-laying following a failure than yearling males (Table 1). However, reproductive performance was also strongly related to whether the field layer height was permanently short or growing tall (ANOVA: field layer height*year: P<0.008 for all above reproductive variables with higher reproductive performance in territories with short compared with tall field layers). Furthermore, date of arrival was correlated with date of breeding (i.e. date of hatching; ANCOVA: year*arrival date: arrival date: F1,496 =91.75, P<0.0001) and reproductive performance (except nest predation risk, see below) was negatively correlated with date of breeding (ANCOVA: year*date of hatching: successful nests: clutch size: F1,305 =49.78, P<0.0001; number fledged: F1,491 =24.40, P<0.0001). Thus, the observed strong patterns of age-specific reproductive performance could at least partly have been caused by age differences in breeding time and type of territory (i.e. permanently short or growing field layer).

Nest Predation Risk and Probability of Re-laying The probability of nest predation was not significantly associated with breeding time as estimated by date of arrival (logistic regression: year*arrival date: arrival date:
21 =1.57, P=0.20). Thus, I analysed the probability of nest predation by including only year, age and territory field layer height in the model. When differences in male age and territory field layer height were taken into account, old males showed a slight (nonsignificant) tendency to have a lower risk of nest predation compared with yearling males (logistic regression: year*male age*field layer height: male age:
21 =3.43, P=0.064; Fig. 2). However, field layer height had a strong significant effect on the

probability of nest predation (field layer height:
21 =55.16, P<0.0001; Fig. 2). Thus, the results of the multiple logistic regression suggest that the observed age difference in nest predation risk (see Table 1) was mainly caused by a corresponding age difference in territory habitat structure. In this population wheatears are mainly singlebrooded, but an average of 20% made a second breeding attempt after a complete failure. Separating the effects of date of nest predation, territory field layer height, year and male age in a logistic regression showed that all variables were significantly associated with the probability of making a second breeding attempt (Table 2). The probability of second breeding attempts was higher for older than younger males and higher for pairs breeding on territories with a permanently short versus a growing field layer (Fig. 2). Pairs were also more likely to make a second attempt when they failed early in the season (Table 2). Second breeding attempts did not differ from first ones in nest predation rate (logistic regression: year*age*field layer height*breeding attempt: breeding attempt:
21 =0.002, P=0.96), but the number of fledged young in successful attempts was significantly lower in second breeding attempts (least square meanSE=4.08 0.27) than in first breeding attempts (least square meanSE=4.800.07; ANOVA: year*breeding attempt* male age*field layer height: breeding attempt effect: F1,508 =6.72, P=0.01).

Age and Reproductive Performance in Successful Pairs I analysed clutch size, number of fledged young and fledging success (i.e. number fledged/number hatched) in relation to male age, territory field layer height and date of

PA } RT: AGE, NEST PREDATION AND TERRITORY QUALITY

50

40 (a)

(b) 216

30 184 20

61 % Re-laying attempts

40 % Depredated nests

200

361

10

30

20

41

80

10 80

0

0

Tall

Short

Short

Tall

Figure 2. The percentage of (a) nests depredated and (b) re-laying attempts following complete nest failure for old (≥2 years; ) and yearling ( ) male northern wheatears nesting in short (<5 cm) versus tall (≥15 cm) field layer heights. Raw data are pooled over years. Numbers above bars are sample sizes.

hatching for pairs that successfully fledged at least one young. Clutch size did not differ in pairs with old or yearling males when variation in hatching date and Table 2. Logistic regression on the probability of making a second breeding attempt in relation to male age, date of nest predation, field layer height class (permanently short versus tall) and year Independents

df

χ2

P

Male age Date of predation Field layer height Year

1 1 1 6

8.64 11.32 11.77 14.01

0.003 0.0008 0.0006 0.03

χ2 values refer to likelihood-ratio chi-square. N=218.

territory field layer height were taken into account (Table 3, Fig. 3). Clutch size was however significantly related to breeding time, and showed a (nonsignificant) tendency to be related to field layer height (Table 3, Fig. 3). Number of fledged young was also associated with field layer height and date of hatching, but male age had an additional independent effect on fledgling production (Table 3, Fig. 3). The above differences in fledgling production were mainly caused by brood reduction because fledging success was significantly associated with male age and field layer height (Table 3, Fig. 3). Thus, most of the observed age differences in fledgling production probably could be attributed to a corresponding age difference in breeding time and territory field layer height, but male age had some additional independent effect on number of fledglings produced.

Table 3. ANCOVAs with clutch size, number of fledged young and fledging success (i.e. number fledged/number hatched) as dependent variables and male age, date of hatching, territory field layer height class (i.e. permanently short versus growing) and year as independent variables

Clutch size: Male age Date of hatching Field layer height Year Error Number fledged: Male age Date of hatching Field layer height Year Error Fledging success: Male age Date of hatching Field layer height Year Error

df

SS

F

P

1 1 1 6 299

0.28 17.52 1.44 4.98 136.36

0.61 38.43 3.16 1.81

0.43 <0.0001 0.076 0.095

1 1 1 6 481

9.08 17.73 47.54 68.31 818.80

5.33 10.42 27.93 6.69

0.021 0.001 <0.0001 <0.0001

1 1 1 6 296

0.58 0.02 1.31 1.35 40.06

4.30 0.17 9.69 1.66

0.039 0.681 0.002 0.131

Only successfully breeding wheatears were analysed. SS=sum of squares. All models were significant (P<0.001).

383

ANIMAL BEHAVIOUR, 62, 2

6.5 106

40

131 69 Clutch size

Hatching date

135 126 35

30

272

57

Tall

Short

1.0

130

100 5 97 86

Short

Tall

Fledging success

211

4

52

6.0

5.5

Tall

Short

6

Number fledged

384

67 57

0.9

52

0.8

Tall

Short

Figure 3. Least square mean±SE reproductive performance of successfully breeding old (≥2 years; ) and yearling ( ) male northern wheatears nesting in short (<5 cm) versus tall (≥15 cm) field layer heights. Numbers above bars are sample sizes. For hatching date, least square means are from models with male age, field layer height and year as independent variables; for other reproductive variables, hatching date was also taken into account (Table 3).

Age and Choice of Breeding Habitat Almost all pairs (99%) chose to breed in habitats that had a short field layer (i.e. <5 cm height) at the time of territory establishment, but in 43% of these territories, the field layer grew tall before fledging of young. Based on the above results of territory field layer height and reproductive performance, one would expect birds to select territories with a permanently short field layer before those with a growing field layer. In fact, earlier-arriving old males chose territories with a permanently short field layer more frequently than later-arriving yearling males (Table 1). However, there was an age difference in the relationship between arrival date and habitat choice as shown by a significant interaction between male arrival date, male age and territory type chosen (i.e. short versus growing field layer) (Table 4). Among old males, territories with a permanently short field layer were on average chosen before those with a growing field layer (meanSE date of arrival (1=1 May)=
0.480.52 and 1.140.62 for short and tall, respectively), while there was no such relationship between date of arrival

and type of territory chosen among yearling males (meanSE=7.380.76 and 7.580.66 for short and tall, respectively).

Differential Survival and Individual Improvements One could argue that the observed cross-sectional age differences in reproductive performance could be due to Table 4. Logistic regression on the probability of choosing a territory with a permanently short field layer (as opposed to one with a growing field layer) in relation to male age, male arrival date and year Independents

df

χ2

P

Male age (A) Male arrival date (B) A*B Year

1 1 1 6

12.27 4.15 4.26 1.32

0.0005 0.042 0.039 0.97

χ2 values refer to likelihood-ratio chi-square. N=669.

PA } RT: AGE, NEST PREDATION AND TERRITORY QUALITY

Table 5. Within-individual comparison of reproductive performance of male wheatears reproducing as yearlings and 2-year-olds Yearling

%Nest predation Number fledged† %In short field layer

2-year-old

N

Mean

SE

Mean

SE

P*

58 47 58

41.4 2.96 46.6

0.41

15.5 4.17 65.5

0.36

0.002 0.037 0.038

*P values refer to chi-square test except for number fledged for which a paired t test was used. †Sample includes males that had their nests depredated.

100 7 % Change in RS

the selective death of poor-quality 1-year-old individuals. Yearling males that survived to the next season had experienced a slightly lower risk of nest predation (23.4%) than those that subsequently died (28.7%), but the relationship between male survival and nest predation was nonsignificant (logistic regression: year*nest predation: N=120 1-year-old males: nest predation:
21 =0.90, P=0.34). However, among successful 1-year-old males (N=79), there was a significant positive relationship between number of fledged young and male survival probability to the next year (logistic regression: year* number fledged: number fledged:
21 =9.65, P=0.002). Males that died before the next season produced an average (SE) of 4.890.16 fledglings, while those that survived produced an average of 5.59 0.19 fledglings. Still, data on males breeding as yearlings and as 2-yearolds show that males improved their breeding success (i.e. in terms of reduced nest predation risk) and fledgling production with increasing age (Table 5). However, most yearlings (69%) shifted territory between years and most shifts were from territories with tall to short field layers (74% of all territory shifts). Within-individual change in breeding success (i.e failed or successful) could attain three values (decrease, no change and increase) and was analysed by means of an ordinal logistic model including year, change in territory field layer height and change in territory (i.e. territory shifts). Within-individual change in breeding success was strongly associated with a corresponding change in territory field layer height (field layer height:
21 =37.52, P<0.0001; Fig. 4) and tended to be associated with territory shifts (males shifting territory improved their success: territory shift:
21 =3.36, P=0.07). Males that shifted their territories from growing field layers to permanently short field layers improved their success, whereas those that shifted from short field layers to tall field layers reduced their success (Fig. 4). Similarly, within-individual change in fledgling production was significantly linked to a corresponding change in territory field layer height (ANOVA: year*change in field layer height*territory shift: field layer effect: F2,38 =11.63, P<0.0001). Among males breeding in the same type of territory (i.e. short or growing) in both years, there was a tendency for an improved reproductive success with age (Fig. 4), but all males that improved their reproductive success had also shifted to a new territory of the same territory field layer height. However, for males that were successful in both years and that bred on the same territory, the production

33

18

50

0

–50

–100

Short to Tall

No change Tall to Short Change in field layer height

Figure 4. Within-individual change in reproductive success (%RS in year 2−%RS in year 1), in relation to within-individual changes in type of territory chosen (e.g. short field layer height in year 1 versus tall field layer height in year 2) for male wheatears breeding as 1and 2-year-olds. Short=territories with a permanently short (<5 cm) field layer; tall=territories with a growing field layer (≥15 cm) at the incubation and nestling stage. Numbers refer to sample sizes.

of fledglings did not significantly improve with increasing age (meanSE difference: year 2 minus year 1=
0.470.44 fledglings; paired t test: t14 =1.05, P=0.29). In fact, the tendency was in the opposite direction; that is, there was a slight reduction in the number of fledglings produced with increasing male age. DISCUSSION Old male wheatears fledged more young than yearling males, because old males experienced a lower nest predation rate, produced more fledglings in successful attempts and more often made second breeding attempts following nest failure (Table 1). These age differences could largely be explained by a corresponding age difference in territory field layer height; older males bred mainly on territories with permanent short vegetation, while yearlings bred on territories with a growing or tall field layer, and field layer height in itself was strongly associated with nest predation risk, fledgling production and probability of re-laying (Figs 2, 3, Tables 2, 3). Longitudinal data strengthens these interpretations by showing that withinindividual improvements in reproductive success with increasing age was mainly a result of a corresponding between-year change of territory field layer height (Fig. 4). However, a smaller, but statistically significant, part of

385

386

ANIMAL BEHAVIOUR, 62, 2

the observed cross-sectional age-specific variation in reproductive performance could not be explained by age differences in territory field layer height (Tables 2, 3). Thus, factors other than territory quality had some additional effect on the observed age-specific reproductive performance.

Age-specific Habitat Selection Studies of habitat selection (see Cody 1985; Newton 1992) and breeding dispersal in birds (Greenwood & Harvey 1982) suggest that old individuals often breed in better habitats than first-time breeders. Even though age differences in territory or nest site quality are commonly inferred as factors causing older individuals to achieve a higher reproductive output than younger ones (e.g. Newton et al. 1981; Afton 1984; Shaw 1986), few studies have explicitly investigated the effects of age-associated differences in breeding habitat on age-specific reproductive performance. Still, as shown in the present study and others (Reese & Kadlec 1985; Møller 1991; Oring et al. 1991; Holmes et al. 1996), age-dependent access to high-quality territories (or other breeding resources) may be a strong factor shaping age-dependent patterns of reproductive success. To understand fully the causes and consequences (e.g. for population dynamics and density-dependent life histories) of age-specific reproductive patterns, one has to understand why older individuals have a competitive advantage in acquiring the best territories. Generally, it is believed to be because of age differences in arrival time on the breeding grounds, dominance and experience (Newton 1992) in combination with a pre-emption mechanism that force young individuals to settle in poorer habitats (Fretwell & Lucas 1970; Pulliam & Danielson 1991). Although the above mechanisms of age-specific habitat selection patterns at least partly apply to wheatears (this study; Conder 1989), there may be other mechanisms as well. This is because my data on order of habitat occupation (as reflected by arrival dates) suggest that yearlings did not seem to be as good as older males at identifying territories with a permanently short field layer (Table 4). All wheatears almost invariably chose territories that had a short field layer upon arrival in April and in May, but in many territories the field layer grew tall during incubation and nestling care. In a dynamic agricultural landscape, where field layer height during spring and summer is to a large extent determined by farming activities, field layer height at the time of territory selection may not be a good predictor of its quality when feeding young. However, old individuals may increase their precision in habitat choice by gathering more accurate information of habitat quality in alternative patches during the breeding and postbreeding period in their first year (e.g. Danchin et al. 1998). In fact yearling male wheatears, and especially those that experience nest predation, have been seen to select future breeding sites during the previous summer (unpublished data). Both field layer height and conspecific reproductive success are used for selecting sites and since both territory field layer height and nest predation risk are

temporally correlated between subsequent years (e.g. 87% of territories with a short field layer in year t
1 had a short field layer in year t), the probability of choosing a good ‘safe’ site in the next year increases (unpublished data). Irrespective of the mechanisms behind age-specific access to high-quality habitats, my study and others strongly suggest that age-specific habitat and territory selection should be kept in mind when analysing the proximate causes of age-specific reproduction.

Territory Quality: Food Limitation and Nest Predation My results suggest that nest predation is a major component determining territory quality of wheatears, an assertion that has often been made for habitat quality in general (e.g. Martin 1988; see also brood parasitism, Robinson et al. 1995). However, food is also an important component of habitat quality as it undeniably affects reproductive performance (e.g. Lack 1954; Martin 1987). In fact, food limitation and nest predation risk may be expected to interact, because guarding and defending nestlings have to be traded against foraging (see Martin 1987), and because food limitation will increase the begging behaviour of nestlings, and consequently, increase their probability of detection (Haskell 1994). A smaller clutch size may result in less begging behaviour, and hence, as a result of nest predation alone, one may expect clutch size and number of fledged young to be lower in risky habitats compared with less risky habitats (Martin 1992; Julliard et al. 1997; Martin et al. 2000). In the case of the wheatear, habitats with a tall field layer are likely to facilitate nest predator activities, since all potential nest predator species (weasel, stoat and adder; see also Tye 1992) prefer habitats with some cover, where their main prey (i.e. voles) tends to be more abundant (e.g. Erlinge 1974). Risky habitats with a tall field layer are also poor foraging habitats for wheatears, since food availability declines with increasing height of the field layer (Conder 1989; Tye 1992). Thus, it seems likely that nest predation and food availability act in concert to cause habitats with a growing field layer to become poor breeding habitats compared with those having a permanently short field layer (see Figs 2, 3, 4).

Other Factors Affecting Age-specific Reproductive Success A part of the cross-sectional age-specific variation in reproductive performance could be explained by differential survival of phenotypes, since yearling males that did not survive to the next breeding season had lower reproductive success than those that survived to an age of 2 years. Thus, it is possible that the residual age-specific variation (i.e. when taking territory field layer height into account; Table 3) was explained by differential survival of phenotypes. However, the longitudinal data displaying within-individual improvements with increasing age of roughly the same magnitude as in the cross-sectional data

PA } RT: AGE, NEST PREDATION AND TERRITORY QUALITY

(cf. Tables 1, 5) suggest that differential survival of phenotypes only had minor effects on the general crosssectional patterns of age-specific reproduction. An age difference in breeding time is another potential factor that may cause age-specific patterns of reproduction since reproductive output typically decreases over the course of the breeding season (e.g. Verhulst & Tinbergen 1991). Although age differences in breeding time probably contributed to the general pattern of agespecific reproductive performance in wheatears, my data suggest it cannot explain all observed residual age-specific variation in reproductive success (Table 3). In an experimental study on shags (Phalacrocorax aristoteles) it was also shown that age differences in breeding time could not explain a corresponding age difference in fledging success (Daunt et al. 1999). Increased local experience and improved skills (e.g. in foraging and breeding activities) also may cause reproductive success to increase with age (Forslund & Pa¨ rt 1995; for good examples see Jansen 1990; Desrochers 1992a, b). However, data on within-individual changes in reproductive performance of wheatears breeding in the same territory in 2 consecutive years did not support these hypotheses, since fledgling production, if anything, decreased with increasing age. In fact, the longitudinal data suggest age-related improvements in reproduction mainly to be governed by a corresponding improvement in the access to high-quality territories (Fig. 4). Similarly, an experimental study of wheatears breeding in the same territory in 2 consecutive years showed strong effects on within-individual changes in reproductive success when field layer height was manipulated in year 2, but no individual improvements when the field layer was kept constant across years (unpublished data). Thus, the possible effects of improved skills in foraging and breeding activities probably were small compared to improved access to high-quality territories.

Conclusion Traditionally, individual improvements in competence, such as foraging skills and breeding experience, have been invoked as possible explanations of agedependent differences in reproductive performance (Forslund & Pa¨ rt 1995). My study underscores the importance of including defended resources, such as territories and nest sites as potential factors, because these may have profound effects on age-specific patterns of reproductive success. Furthermore, it is often assumed that the amount of food available for individuals is the major determinant of age-specific patterns of reproductive performance. However, since nest predation generally is a major determinant of avian reproductive failures (Ricklefs 1969; Martin 1992), and since nest predation risk may affect feeding rates to nestlings (Martin 1992; see also Martin et al. 2000), nest predation risk is bound to be an important factor affecting age-specific reproductive performance. This is partly because nest predation risk may vary with habitat structure and there is an age difference in the access to low-risk habitats. Thus, future studies should include the effects of resource quality and nest predation

(or predation in general) when investigating why old individuals achieve a higher reproductive success than younger ones. Acknowledgments I thank P. Forslund, S. Roos, B. So ¨ derstro ¨ m, A. Qvarnstro ¨ m, S. Ulfstrand and two anonymous referees for constructive criticism on the manuscript. Thanks to M. Amcoff, S. Eriksson, O. Kvarnba¨ ck, M. Hellstro ¨ m, N. Hjort, M. Hoflin, R. Lager, E. Linnarsson, M. Waern and J. Wretenberg, who helped me with the field work. A major part of this paper was written at the University of California, San Diego (UCSD) and I thank T. Price and R. Lande for equipment and facilities. The study was supported by grants from the Swedish Natural Science Research Council (NFR) and Magnus Bergvall’s foundation. The research presented here was evaluated and approved by the Animal Behavior Society’s Animal Care Committee on 19 April 2001. References Afton, A. D. 1984. Influence of age and time on reproductive performance of female lesser scaup. Auk, 101, 255–265. Bernstein, C., Krebs, J. R. & Kacelnik, A. 1991. Distribution of birds amongst habitats: theory and relevance to conservation. In: Bird Population Studies: Relevance to Conservation and Management (Ed. by C. M. Perrins, J-D. Lebreton & G. J. M. Hirons), pp. 317–345. Oxford: Oxford University Press. Charlesworth, B. 1994. Evolution in Age-Structured Populations. Cambridge: Cambridge University Press. Clutton-Brock, T. H. 1988. Reproductive Success. Chicago: University of Chicago Press. Cody, M. L. 1985. Habitat Selection in Birds. New York: Academic Press. Conder, P. 1989. The Wheatear. London: Christoffer Helm. Danchin, E., Boulinier, T. & Massot, M. 1998. Conspecific reproductive success and breeding habitat selection: implications for the study of coloniality. Ecology, 78, 2415–2428. Daunt, F., Wanless, S., Harris, M. P. & Monaghan, P. 1999. Experimental evidence that age-specific reproductive success is independent of environmental effects. Proceedings of the Royal Society of London, Series B, 266, 1489–1493. Desrochers, A. 1992a. Age-related differences in reproduction by European blackbirds: restraint or constraint? Ecology, 73, 1128– 1131. Desrochers, A. 1992. Age and foraging success in European blackbirds: variation between and within individuals. Animal Behaviour, 43, 885–894. Erlinge, S. 1974. Distribution, territoriality and numbers of the weasel Mustela nivalis in relation to prey abundance. Oikos, 25, 308–314. Forslund, P. & Pa¨rt, T. 1995. Age and reproduction in birds: hypotheses and tests. Trends in Ecology and Evolution, 10, 374– 378. Fowler, G. 1995. Stages of age-related reproductive success in birds: simultaneous effects of age, pair-bond duration and reproductive experience. American Zoologist, 35, 318–328. Fretwell, S. D. & Lucas, H. L. 1970. On territorial behaviour and other factors influencing habitat distribution of birds. Acta Biotheoretica, 19, 272–279.

387

388

ANIMAL BEHAVIOUR, 62, 2

Greenwood, P. J. & Harvey, P. H. 1982. The natal and breeding dispersal of birds. Annual Review of Ecology and Systematics, 13, 1–21. Haskell, D. 1994. Experimental evidence that nestling begging behaviour incurs a cost to nest predation. Proceedings of the Royal Society of London, Series B, 257, 161–164. Holmes, R. T., Marra, P. P. & Sherry, T. W. 1996. Habitat specific demography of breeding black-throated blue warblers: implications for population dynamics. Journal of Animal Ecology, 65, 183–195. Jansen, A. 1990. Acquisition of foraging skills by Heron Island Great Barrier Reef Queensland Australia silvereyes Zosterops lateralis chlorocephala. Ibis, 132, 95–101. Jenni, L. & Winkler, R. 1994. Moult and Ageing of European Passerines. London: Academic Press. Julliard, R., McCleery, R. H., Clobert, J. & Perrins, C. M. 1997. Phenotypic adjustment of clutch size due to nest predation in the great tit. Ecology, 78, 394–404. Lack, D. 1954. Natural Regulation of Animal Populations. Oxford: Clarendon. Martin, T. E. 1987. Food as a limit on breeding birds: a life-history perspective. Annual Review of Ecology and Systematics, 18, 453– 487. Martin, T. E. 1988. Area and habitat effects on structure of forest bird assemblages: is nest predation an underlying influence? Ecology, 69, 74–84. Martin, T. E. 1992. Interaction of nest predation and food limitation in reproductive strategies. Current Ornithology, 9, 163–197. Martin, T. E., Martin, P. R., Olson, C. R., Heidinger, B. J. & Fontaine, J. J. 2000. Parental care and clutch sizes in North and South American birds. Science, 287, 1482–1485. Møller, A. P. 1989. Parasites, predators and nest boxes: facts and artefacts in nest box studies of birds. Oikos, 56, 421–423. Møller, A. P. 1991. Clutch size, nest predation, and distribution of avian unequal competitors in a patch environment. Ecology, 72, 1336–1349. Newton, I. 1989. Lifetime Reproduction in Birds. London: Academic Press.

Newton, I. 1992. Experiments on the limitation of bird numbers by territorial behaviour. Biological Reviews, 67, 125–173. Newton, I., Marquiss, M. & Moss, D. 1981. Age and breeding in sparrowhawks. Journal of Animal Ecology, 50, 839–853. Oring, L. W., Reed, J. M., Colwell, M. A., Lank, D. B. & Maxson, S. J. 1991. Factors regulating annual mating success and reproductive success in spotted sandpipers (Actitis macularia). Behavioral Ecology and Sociobiology, 28, 433–442. Pulliam, H. R. & Danielson, B. J. 1991. Sources, sinks and habitat selection: a landscape perspective on population dynamics. American Naturalist, Supplement, 137, 50–66. Reese, K. P. & Kadlec, J. A. 1985. Influence of high density and parental age on habitat selection and reproduction of black-billed magpies. Condor, 87, 96–105. Ricklefs, R. E. 1969. An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology, 9, 1–48. Robinson, S. K., Thompson, F. R. III, Donovan, T. M., Whitehead, D. R. & Faaborg, J. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science, 267, 1987–1990. SAS Institute 1999. JMP. Version 3.2. Cary, North Carolina: SAS Institute. Saether, B-E. 1990. Age-specific variation in reproductive performance in birds. Current Ornithology, 7, 251–283. Shaw, P. 1986. Factors affecting the breeding performance of Antarctic blue-eyed shags Phalacrocorax atriceps. Ornis Scandinavic, 17, 141–150. Sutherland, W. 1996. From Individual Behaviour to Population Ecology. Oxford: Oxford University Press. Svensson, L. 1992. Identification Guide to European Passerines. 4th edn. Stockholm: Ma¨ rstatryck. Tye, A. 1992. Assessment of territory quality and its effects on breeding success in a migrant passerine, the wheatear. Ibis, 134, 273–285. Verhulst, S. & Tinbergen, J. M. 1991. Experimental evidence for a causal relationship between timing and success of reproduction in the great tit Parus m. major. Journal of Animal Ecology, 60, 269–282.