in the Barn Swallow Hirundo rustica

Functional Ecology 2002 A trade-off between clutch size and incubation efficiency Blackwell Science, Ltd in the Barn Swallow Hirundo rustica S. M. ENGSTRAND* and D. M. BRYANT Avian Ecology Unit, Institute of Biological Sciences, University of Stirling, Stirling FK9 4LA, UK Summary 1. To investigate reproductive costs associated specifically with incubation for a singlesex intermittent incubator, clutches of the Barn Swallow (Hirundo rustica Linnaeus) were manipulated in successive trios, involving reduced (n 2), control (n) and enlarged (n + 2) clutch sizes (n is the natural clutch). All manipulated clutches were within the natural range. 2. Manipulations were made shortly after clutch completion and reversed immediately prior to hatching, so the costs of incubation were examined in isolation while the demands of egg-laying and nestling rearing were maintained at natural levels. 3. The efficiency of incubation was dependent on clutch size: the period from laying to hatch was shortest for those incubating reduced (R), and longest for birds incubating enlarged (E) clutches (R: 14 8 ± 0 1 days, n = 28; control (C): 15 3 ± 0 3, n = 24; E: 15 6 ± 0 2 days, n = 28). Also, hatching success was higher amongst R (92%) than either C (85%) or E (81%) clutches. 4. There was no evidence of an incubation cost due to clutch size in terms of parental body condition, provisioning of nestlings, or in the time interval before a second clutch. 5. No difference was found in female nest attendance behaviour or in the decline in egg density (which could reflect water loss) according to our clutch manipulations, indicating that other factors probably underlie the effects observed. 6. It was shown that the number of eggs being incubated affects components of fitness in Barn Swallows, independently of laying and chick-rearing costs. Intraseasonal costs of incubation included prolonged egg development and reduced hatching success among larger clutches but did not affect nestling growth or interclutch intervals. We suggest that an inability to maintain optimal temperature for egg development is likely to underlie the observed responses to clutch size manipulations. Key-words: Costs of incubation, hatching success, nest attendance, swallow. Functional Ecology (2002) Ecological Society Introduction A fundamental assumption of life-history theory is that organisms should allocate resources in ways that will maximize their fitness (Stearns 1989). Limiting resources must be divided between conflicting demands and activities (formalized by the principle of allocation, Sibly & Calow 1986). When applied to iteroparous reproduction, this may involve a trade-off between current reproduction and future reproductive potential (Williams 1966; Charnov & Krebs 1974): an increased reproductive effort at one breeding attempt would result in a reduced residual reproductive value. This could be manifest through a depression of parental Author to whom correspondence should be addressed: E-mail: smb8@st-and.ac.uk *Current address: School of Biology, Bute Building, University of St Andrews, St Andrews KY16 9TS, UK. condition (Moreno, Sanz & Arriero 1999), perhaps lowering the probability of parental survival (Bryant 1979; Tinbergen, van Balen & van Eck 1985) or reducing future breeding performance (Røskaft 1985; Lessells 1986; Hegner & Wingfield 1987; Pettifor, Perrins & McCleery 1988). The viability or the fertility of offspring could be reduced (Gustafsson & Sutherland 1988; Pettifor et al. 1988; Lindén & Møller 1989; Dijkstra et al. 1990; Gustafsson, Qvarnström & Sheldon 1995). Such effects represent the costs of reproduction. They oppose the selective advantages accruing to higher reproductive investment at any single attempt. This balance of selection pressures may result in an optimal level of investment, characterized, in birds, by an optimal clutch and brood size. Because confounding variables may mask such costs under natural conditions, evidence of reproductive trade-offs is most commonly sought by experimental manipulation of reproductive effort. Several such 782

783 Clutch size influences incubation efficiency studies have yielded evidence of trade-offs within the manipulated breeding attempt (Nur 1984; Gustafsson & Sutherland 1988; Pettifor et al. 1988; Lindén & Møller 1989; Dijkstra et al. 1990), during a subsequent breeding attempt within the season (Slagsvold 1982; Røskaft 1985; Lessells 1986), or even carried over to subsequent years (Askenmo 1979; Nur 1984; Røskaft 1985; Reid 1987; Gustafsson & Sutherland 1988; Gustafsson et al. 1995; Nilsson & Svensson 1996; Golet, Irons & Estes 1998; Wernham & Bryant 1998). To examine the costs of increasing reproductive effort fully, manipulation would involve all stages of offspring production. In birds, this should include egg formation and laying, incubation, chick rearing, and often postfledging care. Unfortunately, the costs associated with egg formation, laying and incubation have, until recently, been omitted from such manipulations, the number of offspring generally being manipulated during the final stages of incubation or shortly after hatch (Askenmo 1979; De Steven 1980; Slagsvold 1982; Finke, Milinkovich & Thompson 1987; Gustafsson & Sutherland 1988; Pettifor 1993a,b; Barber & Evans 1995). However, recent studies that do address these early stages suggest that their contribution to the cost of reproduction cannot be ignored (Heaney & Mohaghan 1996; Monaghan & Nager 1997; Reid, Monaghan & Ruxton 2000a; Visser & Lessells 2001). Studies where clutch sizes were manipulated during incubation have generally found that reproductive costs are indeed linked to incubation effort although the nature of the costs may differ (Baltz & Thompson 1988; Coleman & Whittall 1988; Moreno & Carlson 1989; Smith 1989; Moreno et al. 1991; Székely, Karsai & Williams 1994; Heaney & Monaghan 1995, 1996; Siikamaki 1995; Monaghan & Nager 1997; Cichon 2000; Reid et al. 2000a; Visser & Lessells 2001; Ilmonen, Taarna & Hasselquist 2002). We aimed to detect the consequences of incubation costs in isolation. Hence, we manipulated reproductive effort during incubation alone, while leaving earlier and later stages unchanged. We chose to study the Barn Swallow (Hirundo rustica Linnaeus) in which only the female incubates, because male incubation would not confound female responses to manipulations. Nevertheless, this remains a conservative manipulation of costs, because the full cost of producing additional young would include laying eggs and rearing chicks. By monitoring a range of reproductive parameters the existence of a cost associated with clutch size during incubation, acting on either parents or offspring, could be revealed. Although a complete measure of the fitness costs associated with incubation would include survival of offspring and parents and their subsequent reproductive success, we focused on fitness components which could be measured during or immediately following the manipulated breeding attempt. This was because we recognized that effect sizes were likely to be small and hence responses over the short term would be more readily detected than those occurring later. Materials and methods The Barn Swallow, an insectivorous passerine, is a summer migrant in Scotland, arriving to breed in late April and leaving for wintering grounds in September October (Turner & Rose 1989). Swallows nested in farms in mixed arable and grazing land around Stirling, Central Scotland, where birds were loosely colonial, with 1 12, but usually 2 3 nests at each farm (Thompson 1992). In this population, birds are normally double-, rarely triple-brooded within a season. The female alone is responsible for the incubation of a clutch of two to seven (normally four or five) eggs. Initially, sites were visited every 5 days to determine nest usage, then daily during the laying period to establish clutch size. In 1993, both first and second broods were studied; in 1994, time was available to study second broods only. Clutch manipulations were performed in trios, to reduce environmental variation across treatments, where dates of clutch completion of the three nests were within 2 days. Within each trio, nests were assigned randomly to one of three treatment groups: reduced (R), control (C 1 ) and enlarged (E). Two eggs were removed from the R nest and added to the C 1 nest, from which two eggs were removed to the E nest, resulting in a clutch of (n 2) for R nests (n) for C 1 nests and (n + 2) for E nests, where n is the natural clutch size. Because of the high risk of desertion for clutches with fewer than three eggs (Thompson 1992), clutches of four eggs in the R group were reduced by a single egg only and an additional egg donated from outwith the trio to complete the manipulation. In all cases, manipulated clutch size remained within the natural range observed in this population. All eggs were marked with a small number using marker pen at the apex for identification. Clutch manipulations were carried out on day 2 9 ± 0 2 (SEM) of incubation, where d1 was the day of clutch completion, and were reversed towards the end of incubation, on d14 1 ± 0 2, so that each female hatched and subsequently reared her own chicks. When three nests with matched dates of clutch completion could not be found, pairs of R and E nests were formed by simple transfer of two eggs from R to E. Thirty-five first clutches and 35 second clutches were manipulated in 1993 and 52 second clutches were manipulated in 1994. In 1993, 59 unmanipulated clutches (C 0 ) were also studied to provide an additional control for any effects of handling. As no significant differences were found between C 0 nests and those in C 1 for any reproductive parameter, data from both groups were analysed as a single control group (C). Daily weather information was recorded at the Parkhead Meterological Station, situated at Stirling University (3 54 W, 3 8 N), at an altitude of 40 m

784 S. M. Engstrand & D. M. Bryant above sea level, located centrally within the study site. The availability of insect prey to swallows was assessed by the settled volume, transformed using a logarithmic transformation (defined by V t = log 10, volume in m + 1), of insects caught daily in a 12 2 m suction trap next to the meteorological station (Bryant 1978). Reproductive parameters were recorded to estimate short-term reproductive success. It was not possible to record all parameters at each nest as, for example, routine access to some nests was not always feasible, and in some cases it was not possible to capture the female at the appropriate time. Sample sizes for each parameter vary accordingly. FEMALE BODY MASS AND CONDITION Females were captured in the final days of incubation (d14 2 ± 0 4 (SEM), n = 128) using mist nets and measures of mass (g) and structural size (wing, outer tail, inner tail, keel, head-bill and tarsus, Svensson 1970) were recorded. Live body mass in swallows correlates significantly with total body lipid and protein, and has been shown to be improved only trivially as a measure of body condition by standardization for structural size (Thompson 1992). Therefore, body mass was not adjusted for structural size. To account for the diurnal cycle of mass change (Jones 1985; Ward 1992), mass was standardized to the modal time of capture (20.00 h) by assuming an overnight loss of 0 15 g h 1 between 19.00 and 08.00 h (derived from automated balance data, Ward 1992), and a compensatory gain of 0 18 g h 1 throughout the remainder of the day. The standardized values of time-adjusted mass (mass t ) differed from the raw data by an average of 1 4% ± 0 2. Confirming previous data from the study area (Jones 1987; Ward 1992), there was no trend of mass depletion throughout incubation in the swallow (Spearman rank correlation coefficient between the number of days since clutch completion and female mass t, r = 0 051, P = 0 578 for 88 birds at 122 nests). Although the majority of these birds were caught in the final days of incubation, some incidental captures of non-experimental females at various stages of incubation (i.e. day 2 of incubation to hatch) were included in this analysis. No adjustment for the day of incubation was made. Body mass was not adjusted for maximum daily temperature or insect food availability, as these factors accounted for a significant but small proportion of the variation in mass (r 2 = 0 142, P = 0 016) only within the first clutches of 1993, but not later in the season in either 1993 or 1994. Lipid stores were assessed by visible subcutaneous fat deposits, scored on a 5-point scale (1 low, 5 high) at two positions: the posterior edge of the sternum and the tracheal/claviculo-coracoid interclavicular pit, and the two scores summed. Fat score is known to be highly correlated with extractable lipid for hirundines (Bryant & Westerterp 1983). Protein reserves were assessed by the thickness of pectoralis muscle, measured with a portable Krautkramer ultrasonic flaw detector (OSK7) with an Alpha2 Aerotech 10 MHz transducer probe (Agfa-Gevaert, Mortsel, Belgium; Newton 1993) at three locations on the left breast. Pectoralis index (PI) was calculated as the average depth of pectoralis muscle (arbitrary units) per mm keel. PI for four birds could not be measured because of equipment failure. DURATION OF INCUBATION PERIOD This was calculated under the assumption that incubation began on the date of clutch completion (d1). Where hatching of the clutch was spread over 2 or more days, the incubation period was calculated to the date on which the majority of chicks hatched. DECLINE IN EGG DENSITY Eggs of all bird species lose about 16% of their fresh mass during incubation, largely because of loss of water (Rahn & Ar 1974; Drent 1975). The resultant decline in egg density has been used to determine the age of eggs (Lundberg & Väisänen 1979; Furness & Furness 1981). In this study, attempts were made to evaluate the use of the rate of decline in egg density as a measure of embryonic development. We also aimed to determine the extent to which differences related to clutch size may influence egg density through changes in rates of moisture loss from eggs. Egg mass (nearest 0 01 g) was measured at intervals throughout incubation for single eggs of known length (l, mm) and breadth (b, mm, both to nearest 0 1 mm) in 60 second clutches in 1993 and 1994. Density (g cm 2 ) was calculated as mass/volume (V ), using V = klb 2, with k = 0 507 (Hoyt 1979). Linear regressions of egg density on the progress of incubation (days) were computed for 60 nests where 5 ± 2 measurements were made throughout the incubation period. These regressions were significant at α = 0 05 for 93% of nests. NEST ATTENDANCE Incubation schedules were determined for 16 second clutches during 1994, manipulated to form eight pairs of E and R nests. Three thermistor probes were employed to record temperature inside the nest cup and one to sample ambient temperature in the nest environs at intervals of 60 s. Probes were connected to a battery-powered data-logger, situated away from the nest, which prompted sampling and storage of the temperature at each probe every 60 s. Analysis of the temperature record among the eggs enabled identification of the duration of incubation sessions and recess periods. In order to control as much as possible for environmental influences on nest attendance, 20 24 h data were collected simultaneously from each pair of R and E nests, at least 2 days after clutch manipulation.

785 Clutch size influences incubation efficiency Incubation constancy (IC) was calculated from the mean values of session, recess and night-time session durations as the proportion of each 24-h period which the female spent on the nest. HATCHING SUCCESS This was determined as the proportion of the incubated clutch that hatched successfully. Thus, the success of eggs returned to the donor nest contributed to the hatching success of the incubator, rather than the nest in which they hatched. Only in rare cases where clutches were deserted or eggs were damaged in the nest following reversal of the clutch manipulation (when failure to hatch was apparently due to the laying bird, rather than to the incubating female) was hatching failure not attributed to the incubating female. NESTLING GROWTH RATES Any deterioration in the condition of the female through incubation could affect parental ability to feed nestlings (Lifjeld & Slagsvold 1986), so nestling growth rates were used to assess parental ability at this stage. Repeated measurements of nestling mass (g) and wing length (mm) within the periods for which growth is linear for these parameters (d2 10 for mass, d2 20 for wing length, Thompson 1992) were used to calculate the average daily increase in mass (g day 1 ) and wing length (mm day 1 ) within each nest. Rate of change was preferred over a single measure of size or mass, as the latter would be more strongly influenced by hatching asynchrony within the brood or conditions prevailing at the time of any single measure. INTERBROOD INTERVAL This was defined as the time between date of hatch of first clutch and date of laying of the first egg of the second clutch. This was calculated for those birds in 1993 whose first clutches were manipulated and whose second clutches were found. Nests were excluded from this analysis if the first clutch or brood failed prior to fledging. ANALYSIS Statistical analysis followed Zar (1998) and was undertaken using the package SPSS (SPSS Inc., Chicago, Illinois). Normally distributed data were tested using a two factor GLM design, to account for differences between the three clutch manipulation groups and between the three laying periods ; first clutches in 1993, second clutches in 1993, second clutches in 1994. Interactions between the factors were dropped from analyses if P > 0 10. As the natural clutch size laid by a female may reflect female quality, this was included as a covariate in analyses. The Scheirer Ray Hare test (Scheirer, Ray & Hare 1976) was used as a non-parametric equivalent of two-factor analysis of variance where distributions were non-normal or sample sizes were small. As there is a clear directional hypothesis, that the costs of incubating additional eggs should increase in the order Reduced < Control < Enlarged clutches, we employed an ordered heterogeneity (OH) test (Rice & Gaines 1994). This test incorporates both the significance (reflected in the P-value of the non-directional heterogeneity test) and the order of treatment means (reflected in the Spearman s rank correlation coefficient (r s ) between the observed and predicted ordering of treatment means). To obtain the value of r s, treatment group mean residuals from an analysis of variance with laying period as a factor and natural clutch size as a covariate were correlated with the expected order according to the prediction of increasing costs (R < C < E). All OH tests were one-tailed. One-tailed paired t-tests were used to test for a difference between nest attendance parameters for paired enlarged and reduced nests. Data are presented as means ± SEM unless otherwise stated. Medians, used as a measure of central tendency for data with asymmetrical distributions, were calculated to account for tied data (Zar 1998). Results Insect availability was positively, although weakly, related to maximum temperature (partial correlation coefficient r = 0 198, n = 324, P < 0 001, after controlling for date). Despite the general trends of increasing temperature and food availability throughout the season in both years (Fig. 1), making conditions generally more favourable for birds incubating second compared to first clutches (Turner 1982), feeding conditions were particularly poor during the period when second clutches were incubated in 1994 (July August). FEMALE BODY MASS AND CONDITION Female mass did not differ between laying periods (P = 0 62), with natural clutch size (P = 0 15) or with clutch manipulation (Table 1, OH test P = 0 70). However, both fat score, FS (Scheirer Ray Hare, H 2 = 7 03, P = 0 030) and PI (F 2,112 = 12 20, P < 0 0005) differed between laying periods, as birds incubating second clutches in 1993 had larger fat stores and thicker pectoralis muscles (PI 40 8 ± 0 43 arbitrary units mm 1, n = 35, FS median 9 1 (interquartile range, IQR 7 25 10), n = 36) than birds incubating either first clutches in 1993 (PI 39 1 ± 0 47, n = 38, FS 8 0 (IQR 7 9), n = 40) or second clutches in 1994 (PI 37 8 ± 0 43, n = 45, FS 8 0 (IQR 6 10), n = 46). Although both condition indices, FS and PI, differed with treatment group in the direction predicted according to the hypothesis of lower condition for birds incubating enlarged clutches, E < C < R), the effects did not attain significance (OH P = 0 09 for FS, P = 0 09 for PI).

786 S. M. Engstrand & D. M. Bryant between laying periods (F 2,74 = 2 62, P = 0 079) or according to natural clutch size (P = 0 121). However, clutch size manipulation did influence IP in the direction predicted (OH test; P = 0 03), birds with E clutches incubating for longest, those with R clutches the shortest (Table 1, Fig. 2). DECLINE IN EGG DENSITY During the egg-laying period, changes in egg mass were negligible, but from clutch completion to hatch, eggs lost mass at an average rate of 1 1 10 3 g cm 2 day 1 ± 4 6 10 5, n = 60. The rate of decline in egg density did not differ between 1993 and 1994 (second clutches only, P = 0 651) or between clutch manipulation groups (OH test; P = 0 48, Table 1, natural clutch size as covariate P = 0 207). Rates of decline did not correlate with the IP for 54 nests where IP was known accurately (Spearman rank correlation, r = 0 004, P = 0 980), suggesting that the rate of decline in egg density does not reflect the rate of embryonic development in the swallow. Hence, conditions affecting evaporative losses from the nest, such as temperature, humidity and air movements, may be responsible for a greater part of the variation between clutches. Fig. 1. Environmental conditions during the breeding seasons of 1993 ( ) and 1994 ( ): mean ± SEM monthly (a) temperature (24 h maximum, C) and (b) aerial insect availability, measured as V t = log 10 (v + 1), where v is suction trap catch volume (ml). Data for V t early in 1994 are incomplete. The median dates for the start of the incubation period were as follows: 1993: first clutches 1 June; second clutches 23 July; 1994 second clutches: 14 July. DURATION OF INCUBATION PERIOD The mean duration of IP was 15 2 ± 0 1 days (range 12 19 days, n = 80 nests). IP did not differ significantly NEST ATTENDANCE Although all three parameters of nest attendance varied in the direction of more incubation for R clutches, none of the differences attained significance (one-tailed paired t-tests, session length t 7 = 0 05, P = 0 480, recess length t 7 = 0 71, P = 0 249, night session t 7 = 0 31, P = 0 383). The combined measure of incubation constancy IC did not differ between R (83% ± 3) and E (82% ± 3) nests (paired t 7 = 0 62, P = 0 278, Table 1). However, the power of the test to detect small differences in IC between groups is low. With the observed levels of variation and eight birds in Table 1. Reproductive parameters for swallows incubating reduced, control or enlarged clutches. Data are means ± SEM, or medians (calculated accounting for tied data) with interquartile range (IQR) for ordinal data, and sample size. Mass t indicates body mass (g) adjusted to the modal time of capture. Hatching success is determined as the proportion of eggs incubated that hatched successfully excluding complete nest failure. Values marked with an asterisk (*) differ significantly (P < 0 05) from other values marked with an asterisk within the row Reduced Control Enlarged Mass t (g) 21 89 ± 0 26 (42) 21 80 ± 0 21 (40) 21 96 ± 0 23 (40) Fat score (median, IQR) 8 4 (7,10) (42) 8 3 (7,9 75) (40) 7 9 (6,9) (40) Pectoralis thickness index (arbitrary units) 39 5 ± 0 5 (42) 39 1 ± 0 5 (37) 38 7 ± 0 4 (39) Incubation period 14 8* ± 0 1 (28) 15 3 ± 0 3 (24) 15 6* ± 0 2 (28) Decline in egg density ( 10 2 g cm 3 day 1 ) 1 12 ± 0 06 (23) 1 08 ± 0 08 (14) 1 09 ± 0 09 (23) Session (min) 11 6 ± 1 1 (8) 11 5 ± 0 6 (8) Recess (min) 4 5 ± 0 4 (8) 5 1 ± 0 6 (8) Night session (min) 592 0 ± 18 0 (8) 581 5 ± 29 1 (8) % time spent on nest (incubation constancy) 82 13 ± 1 0 (8) 83 3 ± 1 0 (8) Hatching success 98 8* ± 0 8 (42) 91 9 ± 2 3 (53) 86 7* ± 2 4 (44) Nestling mass increment (g day 1 ) 2 28 ± 0 07 (16) 2 17 ± 0 10 (14) 2 33 ± 0 05 (14) Wing length increment (mm day 1 ) 5 12 ± 0 14 (23) 5 16 ± 0 12 (22) 5 34 ± 0 11 (23) Inter-brood interval (day) 34 9 ± 0 7 (9) 33 0 ± 0 9 (22) 36 8 ± 1 8 (10)

787 Clutch size influences incubation efficiency and the clutch size incubated, r s = 0 841, P < 0 001, n = 139, excluding cases of nest failure: for birds incubating three eggs, number of hatchlings was 3 ± 0, n = 37; for four eggs, 3 7 ± 0 1 hatchlings, n = 20; for five eggs, 4 5 ± 0 1 hatchlings, n = 39; for six eggs 5 2 ± 0 2 hatchlings, n = 23; for seven eggs, 6 3 ± 0 2 hatchlings, n = 20). We did not gather evidence on recruitment of young because of the low return rates of offspring to the study area (Thompson 1992), so we cannot determine effects on recruitment in relation to treatments in the long term. Fig. 2. The duration of the incubation period (days) of reduced (white bars), control (shaded) and enlarged (black) clutches from clutch completion to hatch for 80 swallow nests. NESTLING GROWTH RATES Parental ability to rear young, as evidenced by comparing growth in nestling mass, did not differ between treatment groups (OH P = 0 8, natural clutch size P = 0 148, Table 1). As most of these data (38 of 42 nests) were gathered in 1994, no test for an effect of laying period could be undertaken. The rate of increase in wing length did not differ with laying period (P = 0 430) or clutch manipulation (OH test, P = 0 92), with natural clutch size entered as a covariate (P = 0 362). Fig. 3. Hatching success of reduced (white bars), control (shaded) and enlarged (black) clutches by laying period, excluding those nests that failed completely. Bars indicate mean ± SEM. each group, differences of 4 1% in IC (or 65 min per day) would be necessary for power = 0 8. Thus, we cannot wholly rule out small, but real, differences in IC between groups. HATCHING SUCCESS Overall, 10 of 149 nests failed completely, owing to predation, desertion or nest falls. There was no evidence that the proportion of complete nest failures differed between clutch manipulation categories in each laying period (Friedman test χ 2 = 1 463, P > 0 05). Excluding all cases of complete nest failure, hatching success differed according to clutch manipulation (Table 1, OH test; P < 0 0005, natural clutch as covariate P = 0 844). The difference between laying periods (H 2 = 4 8, P = 0 088) and the interaction between laying period and clutch manipulation (H 2 = 8 70, P = 0 069) both approached significance. Hatching success was highest among R clutches in all laying periods, but the contrast was most marked against the reduced success of C and E second broods in 1994 (Fig. 3). Despite this depression in hatching success for enlarged clutches, the total number of hatchlings increased with the number of eggs incubated (Spearman rank correlation between number of hatchlings INTERBROOD INTERVAL Interbrood interval did not differ between treatment groups (OH test, P = 0 11), however, there was a trend towards greater variation (Levene s statistic 3 069, P = 0 058) in this variable for those birds whose first clutches were enlarged compared to other groups (Table 1). Discussion Though the factors constraining reproductive output have long been studied (Lack 1947), it is only relatively recently that the contributions of egg formation, incubation and postfledging care to the constraints on avian clutch size, in addition to that of nestling rearing, have been identified as significant (Moreno & Carlson 1989; Moreno et al. 1991; Heaney & Monaghan 1995, 1996; Monaghan & Nager 1997; Monaghan, Nager & Houston 1998; Thompson, Monaghan & Furness 1998; Bryan & Bryant 1999; Visser & Lessells 2001). Thus, studies that manipulated only the number of chicks after hatch (Askenmo 1979; De Steven 1980; Finke et al. 1987; Hegner & Wingfield 1987; Smith, Källander & Nilsson 1989; Pettifor 1993a,b) may have failed to include the full cost of raising offspring. The additional costs could, for example, help to explain the deviation often observed between the average brood size and the so-called Lack brood size, which results in maximal production of offspring (Lessells 1986; Dijkstra et al. 1990; Partridge 1992). In this study, we found the incubation period to be influenced by clutch manipulation, such that reduced clutches hatched earlier and enlarged clutches later

788 S. M. Engstrand & D. M. Bryant than control clutches. The duration of the incubation period is an important component of fitness as the period for which the clutch is liable to predation or accident is extended. Also, the probability of recruitment decreases for nestlings reared later in the season and the likelihood of producing a second clutch also declines as the season progresses in Barn Swallows (Møller 1990) and other hirundines (Bryant 1979). Hatching success is an obvious component of fitness and was influenced by clutch size manipulation; birds with reduced clutches were more likely to hatch all eggs successfully. This effect was most evident among birds incubating second clutches in 1994. For birds incubating at this time, reserves were perhaps depleted after the first breeding attempt and food availability was lower than for the corresponding period in 1993. It was notable that hatching success among reduced clutches was higher than for both control and enlarged nests. This suggests that, even under natural conditions, factors dependent on clutch size may constrain hatching success, especially under less favourable conditions. In this study, we found no evidence that the cost of incubation was carried over to influence other stages in the breeding cycle or the timing of a subsequent breeding attempt within the season. Few studies have attempted to isolate the costs attributed to the incubation stage alone. Some such studies have compared the performance of birds required to rear additional young with those required to bear the costs of both incubation and rearing (Moreno et al. 1991) and also with full cost birds, stimulated to lay, incubate and rear additional young (Heaney & Monaghan 1995; Visser & Lessells 2001). Recently, other studies have employed a methodology similar to our own, where clutch size manipulations made at the start of incubation were restored either just before (Heaney & Monaghan 1996; Cichon 2000; Reid et al. 2000a) or just after hatch (Ilmonen et al. 2001) to alter effort during incubation only. In each of these studies, the evidence shows that incubation effort has important consequences for fitness, though differences in the nature of the findings between studies suggests that the form of this cost may vary. Female condition during incubation seems to be sacrificed rarely (see also Thomson et al. 1998), though mass loss during breeding may have adaptive value (Freed 1981; Norberg 1981; Moreno 1989), so may not be simply interpreted as evidence for a cost of reproduction. Some studies have found costs of incubation to be evident mainly during the incubation stage (an increase in the duration of the incubation period or a depression in hatching success, this study; Moreno et al. 1991). However, other studies have found little evidence for these immediate costs, but instead have found trade-offs with performance in later stages of reproduction, influencing the number or condition of fledglings (Heaney & Monaghan 1995, 1996; Cichon 2000; Ilmonen et al. 2002) or even female survival (Visser & Lessells 2001). One study found incubation costs to influence both hatching success and fledgling condition (Reid et al. 2000a). It is not yet clear to what extent these differences in response are species-specific or whether they are related to variation in environmental conditions. The mechanism by which manipulation of clutch size influenced both hatching success and the incubation period in this study is not clear. Few studies have reported the effects of clutch manipulation during incubation on patterns of nest attendance. In contrast to our findings, Jones (1987) reported an increase in nest attendance with increasing clutch size for swallows in the same study population. However, these results referred to just two females incubating clutches outwith the natural size range, during afternoons, under generally favourable environmental conditions, and so may not be representative of the circumstances under which constraints normally operate. In the present study, as in previous work (Moreno et al. 1991), we found non-significant trends in the direction of reduced attentiveness for birds incubating large clutches. However, the variation among individuals suggests that much larger sample sizes would be necessary to detect small changes in incubation constancy. In the absence of a significant difference in attendance schedules in this study, we conclude that changes in the nest microclimate related to female behaviour while on the nest were more likely to be responsible for reductions in hatching success and extension of the incubation period than periodic neglect of clutches. Physical factors dependent on clutch size can influence both the temperature and humidity of the nest. Maintenance of correct nest humidity is important in controlling rates of water loss from eggs during incubation (Walsberg & Schmidt 1992) and can have implications for hatching success (Packard & Packard 1993). In warm and arid conditions in Spain, Reid et al. (2000a) reported that starling Sturnus vulgaris eggs incubated in enlarged clutches lost a greater proportion of their initial mass than eggs in control clutches. As mass loss largely reflects water loss (Rahn & Ar 1974), this difference was attributed to an increase in the rate of water loss in enlarged clutches. Reid et al. (2000a) suggested that this could underlie the reductions in hatching success which were observed among enlarged clutches. However, in the present study, no difference in mass loss was detected between eggs incubated in different groups. We suspect that the lower temperatures at our study site allowed more effective control of nest humidity. We suggest that differences in egg temperature were the main cause of the differences in hatching success and incubation period observed in our study. If the clutch is too large for all eggs to be maintained in simultaneous full-contact with the brood patch, peripheral eggs may be exposed to periods of cooling before the clutch is turned and re-heated, leading to

789 Clutch size influences incubation efficiency reductions in the rate of development (Haftorn 1983) and increasing the risk of developmental abnormality (Webb 1987). The increased insulation provided by the eggs in a large clutch, however, can reduce rates of cooling from an unattended clutch as well as rates of re-heating as the female resumes incubation (Reid et al. 2000a). In warm conditions these conflicting factors can combine to produce a positive association between average egg temperatures and clutch size (Reid et al. 2000a). However, the relationship between egg temperature and clutch size in cooler conditions, such as prevailed in this study, has not been studied. As prolonged incubation periods are normally associated with reductions, rather than increases, in egg temperature during development (Webb 1987), we suspect that average egg temperature was lower, but perhaps more variable, in enlarged compared to reduced clutches. Future studies could investigate the influence of clutch size on rates of egg turning behaviour and egg temperature under a range of environmental conditions. This study contributes to recent evidence that incubation may itself impose reproductive costs on birds. This is consistent with energy expenditure in the field during incubation not being reduced compared to other stages in the annual cycle among uni-parental incubators (Tatner & Bryant 1993; Williams 1996), and being influenced by clutch size (Moreno et al. 1991; Moreno & Sanz 1994; Engstrand, Ward & Bryant 2002). The risk of predation, a major selection pressure shaping incubation systems, can be influenced by clutch size if nest attendance patterns during incubation or its duration are affected (Conway & Martin 2000). Manipulation of clutch size (Moreno et al. 1991; Thomson et al. 1998; Reid et al. 2000a; Heaney & Monaghan 1995, 1996; Visser & Lessells 2001; this study) or even modest manipulations in the nest microclimate during incubation (Bryan & Bryant 1999; Reid et al. 2000b) can also influence fitness in a significant way. It is therefore important that future studies considering the evolution of clutch size recognize the specific costs of incubation as well as those arising from other stages of the reproductive cycle. Acknowledgements We are grateful to the landowners and farmers in the Stirling area who granted permission to work on their land and to Mike Taylor who assisted with fieldwork. The study was carried out while S.E. held a NERC studentship. References Askenmo, C. (1979) Reproductive effort and return rate of male pied flycatchers. American Naturalist 114, 748 753. Baltz, M.E. & Thompson, C.F. (1988) Successful incubation of experimentally enlarged clutches by house wrens. Wilson Bulletin 100, 70 79. Barber, C.A. & Evans, R.M. (1995) Clutch-size manipulations in the yellow-headed blackbird: a test of the individual optimisation hypothesis. Condor 97, 352 360. Bryan, S.M. & Bryant, D.M. (1999) Nest-box heating reveals an energetic constraint on incubation in the great tit Parus major. Proceedings of the Royal Society of London Series. B 266, 157 162. Bryant, D.M. (1978) Environmental influences on growth and survival of nestling house martins Delichon urbica. Ibis 120, 271 283. Bryant, D.M. (1979) Reproductive costs in the house martin, Delichon urbica. Journal of Animal Ecology 48, 655 675. Bryant, D.M. & Westerterp, K.R. (1983) Short term variability in energy turnover by breeding house martins Delichon urbica, a study using doubly labelled water D 2 18 O. Journal of Animal Ecology 52, 525 543. Charnov, E.L. & Krebs, J.R. (1974) On clutch size and fitness. Ibis 116, 217 219. Cichon, M. (2000) Costs of incubation and immunocompetence in the collared flycatcher. Oecologia 125, 453 457. Coleman, R.M. & Whittall, R.D. (1988) Clutch size and the cost of incubation in the bengalese finch (Lonchura striata var. domestica). Behavioural Ecology and Sociobiology 23, 367 372. Conway, C.J. & Martin, T.E. (2000) Evolution of passerine incubation behaviour: influence of food, temperature and nest predation. Evolution 54, 670 685. De Steven, D. (1980) Clutch size, breeding success and parental survival in the tree swallow (Iridoprocne bicolor). Evolution 34, 278 291. Dijkstra, C., Bult, A., Bijlsma, S., Daan, S., Meijer, T. & Zijlstra, M. (1990) Brood size manipulations in the kestrel (Falco tinnunculus): effects on offspring and parent survival. Journal of Animal Ecology 59, 269 s285. Drent, R.H. (1975) Incubation. Avian Biology, Vol. 5 (eds D.S. Farner & J.R. King), pp. 333 420. Academic Press, London and New York. Engstrand, S.M., Ward, S. & Bryant, D.M. (2002) Variable energetic responses to clutch size manipulations in whitethroated dippers Cinclus cinclus. Journal of Avian Biology 33, 371 379. Finke, M.A., Milinkovich, D.J. & Thompson, C.F. (1987) Evolution of clutch size: an experimental test in the house wren (Troglodytes aedon). Journal of Animal Ecology 56, 99 114. Freed, L.A. (1981) Loss of mass in breeding wrens: stress or adaptation? Ecology 62, 1179 1186. Furness, R.W. & Furness, B.L. (1981) A technique for estimating the hatching dates of eggs of unknown laying date. Ibis 123, 98 102. Golet, G.H. & Irons, D.B. & Estes, J.A. (1998) Survival costs of chick rearing in black-legged kittiwakes. Journal of Animal Ecology 67, 827 841. Gustafsson, L. & Sutherland, W.J. (1988) The costs of reproduction in the collared flycatcher Ficedula albicollis. Nature 335, 813 815. Gustafsson, L., Qvarnström, A. & Sheldon, B.C. (1995) Trade-offs between life-history traits and a secondary sexual character in male collared flycatchers. Nature 375, 311 313. Haftorn, S. (1983) Egg temperature during incubation in the great tit (Parus major), in relation to ambient temperature, time of day, and other factors. Fauna norvica Series C Cinclus 6, 2 38. Heaney, V. & Monaghan, P. (1995) A within-clutch trade-off between egg production and rearing in birds. Proceedings of the Royal Society of London Series B 261, 361 365. Heaney, V. & Monaghan, P. (1996) Optimal allocation of effort between reproductive phases: the trade-off between incubation costs and subsequent brood rearing capacity.

790 S. M. Engstrand & D. M. Bryant Proceedings of the Royal Society of London Series B 26, 1719 1724. Hegner, R.E. & Wingfield, J.C. (1987) Effects of brood size manipulations on parental investment, breeding success and reproductive endocrinology of house sparrows. Auk 104, 470 480. Hoyt, D.F. (1979) Practical methods of estimating volume and fresh weight of bird eggs. Auk 96, 73 77. Ilmonen, P., Taarna, T. & Hasselquist, D. (2002) Are incubation costs in female pied flycatchers expressed in humoral immune responsiveness or breeding success. Oecologia 130, 199 204. Jones, G. (1985) Parent: offspring resource allocation strategies in birds: studies on swallows (Hirundinidae). PhD Thesis, University of Stirling, Stirling. Jones, G. (1987) Time and energy constraints during incubation in free-living swallows (Hirundo rustica): an experimental study using precision electronic balances. Journal of Animal Ecology 56, 229 245. Lack, D. (1947) The significance of clutch size. Ibis 89, 302 352. Lessells, C.M. (1986) Brood size in Canada geese: a manipulation experiment. Journal of Animal Ecology 55, 669 689. Lifjeld, J.T. & Slagsvold, T. (1986) The function of courtship feeding during incubation in the pied flycatcher Ficedula hypoleuca. Animal Behaviour 34, 1441 1453. Lindén, M. & Møller, A.P. (1989) Cost of reproduction and covariation of life history traits in birds. Trends in Ecology and Evolution 4, 367 371. Lundberg, C.A. & Väisänen, R.A. (1979) Selective correlation of egg size with chick mortality in the black-headed gull (Larus ridibundus). Condor 81, 146 156. Møller, A.P. (1990) Male tail length and female mate choice in the monogamous swallow Hirundo rustica. Animal Behaviour 39, 458 465. Monaghan, P. & Nager, R.G. (1997) Why don t birds lay more eggs? Trends in Ecology and Evolution 12, 270 274. Monaghan, P., Nager, R.G., Houston, D.C. (1998) The price of eggs: increased investment in egg production reduces the offspring rearing capacity of parents. Proceedings of the Royal Society of London Series B 265, 1731 1735. Moreno, J. (1989) Strategies of mass change in breeding birds. Biological Journal of the Linnaean Society 37, 297 310. Moreno, J. & Carlson, A. (1989) Clutch size and the costs of incubation in the pied flycatcher Ficedula hypoleuca. Ornis Scandinavica 20, 123 128. Moreno, J., Gustafsson, L., Carlson, A. & Pärt, T. (1991) The cost of incubation in relation to clutch-size in the collared flycatcher Ficedula albicollois. Ibis 133, 186 193. Moreno, J. & Sanz, J.J. (1994) The relationship between the energy expenditure during incubation and clutch size in the Pied Flycatcher Ficedula hypoleuca. Journal of Avian Biology 25, 125 130. Moreno, J., Sanz, J.J. & Arriero, E. (1999) Reproductive effort and T-lymphocyte cell-mediated immunocompetence in female pied flycatchers Ficedula hypoleuca. Proceedings of the Royal Society of London Series B 266, 1105 1109. Newton, S.F. (1993) Body condition of a small passerine bird: ultrasonic assessment and significance in overwinter survival. Journal of Zoology. 229, 561 580. Nilsson, J.A. & Svensson, E. (1996) The cost of reproduction: a new link between current reproductive effort and future reproductive success. Proceedings of the Royal Society of London Series B 263, 711 714. Norberg, R.A. (1981) Temporary weight decrease in breeding birds may result in more fledged young. American Naturalist 118, 838 850. Nur, N. (1984) The consequences of brood size for breeding blue tits. I. Adult survival, weight change and the cost of reproduction. Journal of Animal Ecology 53, 479 496. Packard, M.J. & Packard, G.C. (1993) Water-loss from eggs of domestic fowl and calcium status of hatchlings. Journal of Comparative Physiology 163, 327 331. Partridge, L. (1992) Measuring reproductive costs. Trends in Ecology and Evolution 7, 99 100. Pettifor, R.A. (1993a) Brood manipulation experiments 1. The numbers of offspring surviving per nest in blue tits (Parus caeruleus). Journal of Animal Ecology 63, 131 144. Pettifor, R.A. (1993b) Brood manipulation experiments 2. A cost of reproduction in blue tits (Parus caeruleus). Journal of Animal Ecology 63, 145 159. Pettifor, R.A., Perrins, C.M. & McCleery, R.H. (1988) Individual optimisation of clutch size in great tits. Nature 336, 160 162. Rahn, H. & Ar, A. (1974) The avian egg: incubation time and water loss. Condor 76, 147 152. Reid, W.V. (1987) The cost of reproduction in the glaucouswinged gull. Oecologia 74, 458 467. Reid, J.M., Monaghan, P. & Ruxton, G.D. (2000a) The consequences of clutch size for incubation conditions and hatching success in starlings. Functional Ecology 14, 560 565. Reid, J.M., Monaghan, P. & Ruxton, G.D. (2000b) Resource allocation between reproductive phases: the importance of thermal conditions in determining the cost of incubation. Proceedings of the Royal Society of London Series B 267, 37 41. Rice, W.R. & Gaines, S.D. (1994) Extending nondirectional heterogeneity tests to evaluate simply ordered alternative hypotheses. Proceedings of the National Academy of Science, USA 91, 225 226. Røskaft, E. (1985) The effect of enlarged brood size on the future reproductive potential of the rook. Journal of Animal Ecology 54, 255 260. Scheirer, C.J., Ray, W.S. & Hare, N. (1976) The analysis of ranked data derived from completely randomised factorial designs. Biometrics 32, 429 434. Sibly, R.M. & Calow, P. (1986) Physiological Ecology of Animals: an Evolutionary Approach. Blackwell Scientific Publications, Oxford. Siikamaki, P. (1995) Are large clutches costly to incubate? The case of the pied flycatcher. Journal of Avian Biology 26, 76 80. Slagsvold, T. (1982) Clutch size, nest size and hatching asynchrony in birds: experiments with the fieldfare (Turdus pilaris). Ecology 63, 1389 1399. Smith, H.G. (1989) Larger clutches take longer to incubate. Ornis Scandinavica 20, 156 158. Smith, H.G., Källander, H. & Nilsson, J.Å. (1989) The trade-off between offspring number and quality in the great tit Parus major. Journal of Animal Ecology 58, 383 401. Stearns, S.C. (1989) Trade-offs in life-history evolution. Functional Ecology 3, 259 268. Svensson, L. (1970) Identification Guide to European Passerines. Svensson, Stockholm. Székely, T., Karsai, I. & Williams, T.D. (1994) Determination of clutch size in the Kentish plover Charadrius alexandrinus. Ibis 136, 341 348. Tatner, P. & Bryant, D.M. (1993) Interspecific variation in daily energy expenditure during avian incubation. Journal of Zoology 231, 215 232. Thompson, M.L.P. (1992) Reproductive success and survival of swallows (Hirundo rustica): effects of age and body condition. PhD Thesis, University of Stirling, Stirling. Thomson, D., Monaghan, P. & Furness, R.W. (1998) The demands of incubation and avian clutch size. Biological Reviews of the Cambridge Philosophical Society 73, 293 304. Tinbergen, J.M. & van Balen, J.H. & van Eck, H.M. (1985)

791 Clutch size influences incubation efficiency Density-dependent survival in an isolated great tit population: Kluyver s data reanalysed. Ardea 73, 38 48. Turner, A.K. (1982) Timing of laying by swallows (Hirundo rustica) and sand martins (Riparia riparia). Journal of Animal Ecology 51, 29 46. Turner, A.K. & Rose, C. (1989) A Handbook of the Swallows and Martins of the World. Christopher Helm, London. Visser, M.E. & Lessells, C.M. (2001) The costs of egg production and incubation in great tits (Parus major). Proceedings of the Royal Society of London Series B 268, 1271 1277. Walsberg, G.E. & Schmidt, C.A. (1992) Effects of variable humidity on embryonic development and hatching success of mourning doves. Auk 109, 309 314. Ward, S. (1992) Energetics of laying and incubation in birds: studies of swallows Hirundo rustica, dippers Cinclus cinclus and Japanese quail Coturnix coturnix. PhD Thesis, University of Stirling, Stirling. Webb, D.R. (1987) Thermal tolerance of avian embryos: a review. Condor 89, 874 898. Wernham, C.V. & Bryant, D.M. (1998) An experimental study of reduced parental effort and future reproductive success in the puffin Fratercula arctica. Journal of Animal Ecology 67, 25 40. Williams, G.C. (1966) Natural selection, the costs of reproduction, and a refinement of Lack s principle. American Naturalist 100, 687 690. Williams, J.B. (1996) Energetics of avian incubation. Avian Energetics and Nutritional Ecology (ed. C. Carey), pp. 375 416. Chapman & Hall, New York. Zar, J.H. (1998) Biostatistical Analysis. Prentice Hall International, Englewood Cliffs, NJ. Received 7 February 2002; revised 31 May 2002; accepted 7 June 2002