Growth and food requirement exibility in captive chicks of the European barn owl (Tyto alba)

J. Zool., Lond. (1998) 245, 137±145 # 1998 The Zoological Society of London Printed in the United Kingdom Growth and food requirement exibility in captive chicks of the European barn owl (Tyto alba) Centre d'ecologie et Physiologie EnergeÂtiques, Centre National de la Recherche Scienti que, F-6787 Strasbourg Cedex 2, France; e-mail: handrich@c-strasbourg.fr (Accepted 3 September 1997) Abstract The growth and the food requirements of the European barn owl (Tyto alba) were studied in three groups of captive chicks. One group of chicks, raised in ad libitum food conditions by their parents, was used to measure body mass and linear growth of a number of structural body components at their characteristic growth rate. A second group of chicks, hand-raised with food ad libitum until edging, was used to measure food requirement at the characteristic growth rate. A third group, hand-raised with restricted feeding, was used to specify the plasticity of growth and food requirements still compatible with a normal development at edging. Chicks of both sexes of the rst two groups experienced an overshoot in body mass (maximum of 391 g) when approximately 4 days old, followed by a decrease (to a mass of 314 g) until edging at 6 days old, giving a mass distribution with age in the form of a bell-shaped curve. The daily food intake also showed a bell-shaped curve with a peak value of 8 1 g of fresh mice/day when the body mass was maximum. The decrease of food intake preceding edging was spontaneous. The average daily food intake was 67 17 g/day between the ages of 2 and 6 days. In the restricted feeding group, however, the same timing for linear growth and edging was achieved with a 17% reduction of daily food intake. For the same edging body mass as the ad libitum fed group, there was a attening of the body mass curve. The signi cance of the overshoot in body mass and food intake is discussed in terms of the chick's and brood's advantages and in terms of parental investment in this species with hatching asynchrony. Key words: Tyto alba, raptors, hatching asynchrony, brood requirement, logistic model INTRODUCTION In numerous bird species, the postnatal growth shows a bell-shaped curve of body mass characterized by a peak body mass heavier than that of an adult and a subsequent decrease prior to edging (review in Ricklefs, 1968). This peak has been interpreted as a fat accumulation (Lacombe, Bird & Hibbard, 1994) that must be reduced prior to edging by a reduction of food intake (Ricklefs, 1968; Ricklefs & Schew, 1994). The accumulation of body reserves has been considered as an adaptation to stochastic variation in foraging experienced by parents (the energy insurance hypothesis). However, there is no direct data to assert that the starvation period prior to edging is not an endogenous phenomenon independent of the chicks' food availability. Hatching asynchrony, which is a breeding strategy often observed in raptors, is generally interpreted as an adaptation to stochastic variation in food availability (brood reduction hypothesis; Lack, 1968). Different species also show the characteristic bell-shaped growth curve. In the American kestrel Falco sparverius (Lacombe et al., 1994) it was demonstrated that a 1% reduction of energy intake causes a disappearance of the overshoot of body mass without any consequence for body mass or body lengths at edging time. In raptors with a large clutch size (e.g. barn, snowy or short-eared owl) and asynchronous hatching, we might expect the effects of uctuations in food availability to be intricate since it could affect each chick according to birth rank. In these situations, the possibility of modifying the growth could be an advantage for chick survival and/or for parental effort. In the barn owl (Tyto alba), the edglings (yearly mean range 1.5±5.; Taylor, 1994) leave the nest after more than two months of growth. Interestingly, edging occurs about 2 days after attainment of peak body mass (at 4 days of age: Pickwell, 1948; SchoÈnfeld & Girbig, 1975; Baudvin, 1986; Wilson, Wilson & Durkin, 1987; Langford & Taylor, 1992; Taylor, 1994). Barn

138 owls are also known for yearly uctuations of breeding success (number of breeding pairs, second clutches, clutch size) and nest mortality (Baudvin, 1978; Wilson, Wilson & Durkin, 1986; Taylor, 1994). We investigated the food requirements of captive chicks during growth and the consequences of food restriction on their growth parameters and growth curve shape. Studies in captivity allow measuring growth at its characteristic rate, which is possible only when food availability is high (Ricklefs, 1973; Lenton, 1984). To understand the in uence of food intake on chick growth, we compared groups of chicks raised with food ad libitum with another group restricted in feeding. The major objectives of this study were: (i) to study the chick growth and the corresponding food intake in situations of high food availability; (ii) to learn if the decrease in body mass observed before edging is spontaneous or results from a lessening of food availability; and (iii) to determine the in uence of restricted feeding on chick development and the impact of this food restriction on the foraging investment of the parents. MATERIALS AND METHODS This study was performed on 3 barn owl chicks from a breeding programme conducted since 1985 in the laboratory. The breeding programme was begun with 3 chicks born in captivity from road-injured birds kept in a rescue centre. Whenever possible, new individuals of the same origin were added. Each pair of owls was housed in outside aviaries where a nest box was installed. The pairs bred at least once each year. According to the type of rearing, we have constituted 3 groups of chicks as explained below. Groups of birds The growth of 5 chicks was followed for a rst group. Subgroups were constituted after edging, based on the possibility of getting correct weighings (see below). We selected 1 females out of 21 and 1 males among 29. In this rst group (Control group, n = 2), chicks were raised by their parents. The 1 males came from 5 different broods and 3 pairs of adults. The 1 females came from 6 different broods and 4 pairs of adults. Brooding occurred between March and July in outside aviaries (5 6 4 6 2.5 m), the temperature ranging from 1±28 8C. Since the chicks were raised by their parents, they were considered wild. This group was used as the Control group. In edging barn owls, sexing by external features is not always reliable (Taylor, 1994; Handrich, pers. obs.). Thus, when necessary, sexing was made by laparoscopic examination of gonads under halothane anaesthesia. The procedure complied with current laws of French authorities: Authorization of the MinisteÁre de l'agriculture et de la ForeÃt followed by acceptance of the research programme by the MinisteÁre de l'environnement. In the second group (Ad libitum group, n = 5), the chicks (2 males and 3 females) came from 3 different broods and 3 pairs of adults. Examination of the daily food intake began when the chicks were 2 days old, the age of thermal emancipation (Taylor, 1994). It was then possible to house the chicks separately and alone in open-topped nest boxes (25 6 25 6 5 cm) without the need of female constant care. The boxes were put in a climatic chamber that maintained the natural photoperiod (experiment conducted in June) and an ambient temperature of 25 1 8C. The thermal neutrality zone of the barn owl is 22.5±33 8C (Johnson, 1974; Edwards, 1987). In the third group (Restricted group, n = 5), the chicks (3 males and 2 females) came from 3 different broods and 3 pairs of adults. The chicks were raised under the same conditions as the Ad libitum group except that they were raised alone from their 12th day. This corresponds to the time when the female starts to forage during bad food years (Taylor, 1994). To limit potential heat loss, cotton was added inside the nest boxes. Feeding Chicks of the Control group were fed by their parents with laboratory mice (Mus musculus). Daily, live mice were put in the aviary at 16:. At the same hour, we checked food availability in the nest boxes so that we could adjust the number of mice given. Chicks of the Ad libitum group were also provided with freshly killed mice at 16: each day. At 2 days old, when the experiment started, they were able to eat an entire mouse on their own (Taylor, 1994). Since mice were always available when food was added, they were considered fed ad libitum. Chicks of the Restricted group were fed by hand with pieces of fresh mice and thus the experiment could be begun on 12-day-old chicks. In contrast with the Ad libitum group, the meals were given 3 times daily during daytime only (9:, 12:3 and 16:). The food was ad libitum during meals only and not between. During night-time (16:±9:), the chicks were given no food. Since the meal-to-pellet interval for owls is 1±13 h (Duke, Evanson & Jegers, 1976), and is certainly less for chicks fed with torn mice, the feeding was not ad libitum but restricted. At the age of 2 days, the chicks could feed by themselves and eat the food stored inside the nest. Prior to 2 days old, the chicks need to be fed by the female. As a consequence, chicks younger than 2 days have no food available when the female forages. They will be fed only when the female stays in the nest to rest during daytime (Durant, pers. obs.). This situation was simulated in the Restricted group.

Growth and food intake in the European barn owl 139 Measurements Body mass The chicks were weighed to the nearest.1 g using a Sartorius scale every day before the meal at 16: for the Ad libitum group and the Restricted group but only once every 2 days at the same hour for the Control group. We veri ed by checking the nest box before the weighing that the last pellet was made. When no pellet was found, the chick's belly was gently felt. This super- cial examination was done with great care and no apparent consequences were observed for the birds. Unreliable weighings (weighing with either pellet or meal in stomach) were discarded. Body size Body length (humerus, tarso-metatarsus, tibia, rectrix and remex) were measured with callipers to the nearest mm on the Ad libitum and Control groups during the afternoon weighing. Feather length corresponded to the measurement of feathers plus the remnant of the sheath which sticks out of the epidermis. Remex length was taken to be the length of the tenth primary remex (P1). Rectrix length was the length of the sixth rectrix from the right (R1). Growth curves and statistical analysis To compare the growth data among groups and chicks, we used the graphical method of Ricklefs (1967). A logistic equation was tted to both body mass and linear growth data for each individual (SigmaPlot 5. software from Jandel Scienti c). Even if the logistic equation does not describe a bell-shaped curve, we used it for comparison purposes with other studies. Since the body mass curve is bell-shaped, and because the logistic equation describes a sigmoid curve, we divided the data into 2 parts. The cut was made at the peak of body mass, the 2 delimited parts approximating to a sigmoid form well described by the logistic equation: Y(t) = A [1+e -K (t ± ti) ] -1 where Y(t) = body mass (g) or length (mm) of the chick at age t (days), A = asymptotic body mass (g) or component length (mm), K = growth constant (1/day), ti = in ection point of the growth curve (days). The comparisons were made with the 3 parameters obtained with the tting (A, K and ti) for each individual. The mean values of individual parameters and their standard deviations were then calculated. Statistical analyses were performed with non-parametric tests because of the small group sizes. The Mann±Whitney U-test (MWU-test) was used for the comparisons between sexes, the Kruskal±Wallis test to compare the different groups and the Wilcoxon test to compare changes in the same group. Table 1. Differences between males and females in the Control group. The tted parameters were obtained by the logistic equation for each individual for different growth variables. A is the asymptote (in g or mm), K is the growth constant (in day -1 ) and ti the in ection point. The statistical analysis was performed with the Mann±Whitney U-test using the individual data. No statistical difference was found between sex for each criterion (P>.1). The values are means S.D. Males Females Growth variables Parameter (n = 1) (n = 1) Body mass A 387.32 15.4 396.81 19.86 K.175.13.18.12 ti 17.58 1.11 16.78 1.51 Humerus length A 92.48 4.37 92.85 5.97 K.157.48.154.39 ti 13.13 2.1 13.45 3.22 Tarso-metatarsus A 75.35 4.17 74.79 6.68 length K.15.6.14.4 ti 13.69 2.25 13.89 2.85 Tibia length A 97.54 2.74 97.75 5.67 K.145.31.142.3 ti 12.4 1.29 12.11 1.77 Rectrix length (R1) A 19.88 2.1 13.79 11.1 K.135.9.142.211 ti 39.92 1.59 38.98 2.23 Remex length (P1) A 23.27 7.4 2.55 5.18 K.98.3.99.4 ti 4.5.71 38.98 2.23 RESULTS Growth in the Control group Body mass differences between sexes Although females tended to have an asymptotic body mass greater than males (Table 1 and Fig. 1), the chicks exhibited no signi cant sex differences for the tted parameters (MWU-test, P>.5). Data for males and females were therefore pooled (Table 2 and curve in Fig. 1). Body mass growth The chicks were hatched with a body mass of 13.9 1.4 g. The gain of mass was almost linear between the 8th and 27th day (growth rate of 14.3 1.3 g/day). Near the 4th day, the body mass reached a maximum value of 391. 11.5 g (Table 2). Chicks grew from 1 to 9% of their maximum weight in 25 days (the t 1±9 index of Ricklefs, 1968). Between the 4th day and edging (6 to 65th day; Pickwell, 1948; Taylor, 1994), the birds lost weight. Fledging body mass (314.4 26.5 g) was signi cantly less (Wilcoxon test, P<.1) than the maximum weight. Linear growth Bone and feather growth (Table 1) exhibited no

14 4 a b 3 Body mass, (g) g 2 1 Males Females 1 2 3 4 4 5 6 7 8 9 Age, (days) Fig. 1. Body mass (a) from hatching to maximum weight and (b) from maximum weight to 9 days of age (data for 1 males and 1 females of the Control group and the logistic curve is for the sex-pooled data). The cross-hatched part of (b) represents data already shown in (a). Table 2. Fitted parameters obtained by the logistic equation for different growth variables in the Control group (n = 2). The equation was tted to each individual chick. A is the asymptote, K the growth constant and ti the in ection point. The values are means S.D. Variable Parameter A (g, mm) K (day -1 ) ti (days) Body mass 39.6 11.5.177.3 17.18 1.35 Humerus length 91.59.65.15.3 12.41.18 Tarso-metatarsus length 73.35.61.132.4 12.64.22 Tibia length 96.57.61.14.3 11.37.16 Rectrix length (R1) 19.5.6.14.3 4.33.17 Remex length (P1) 21.9.83.99.1 4.37.15 signi cant difference in the tted parameters between sexes (MWU-test, P>.1 for the bones and P>.1 for the feathers). Therefore, data for males and females were pooled (Table 2 and Fig. 2). Humerus, tibia and tarso-metatarsus lengths were 12. 6.4, 15.2 5.4 and 11.6 4.4 mm, respectively, at hatching. The three bones had nearly nished their growth when the chicks were at their maximum weight (98, 98 and 95% for humerus, tibia and tarso-metatarsus, respectively) and had reached full size at edging. The remex (P1) began to grow out of the skin near the 7th day and continued until the chicks left the nest. At edging, the remex had reached only 8% of the asymptotic length. The rectrix (R1) began its growth later than the remex at around the 17th day but grew more quickly, with a length equal to 88% of the asymptotic value at edging. At the maximum of body mass, the two types of feathers had reached only 47% (P1) and 5% (R1) of their nal length. Food intake and effect of diet on growth rate There was also no signi cant difference for the tted parameters between the Control and Ad libitum groups (MWU-test, P>.1 in Table 3). We therefore considered the Ad libitum group instead of the Control group, for which we did not have the measurement of food intake, as the reference for the relationship between food intake and the characteristic growth. The mean daily food intake was calculated over 4-day intervals (Fig. 3a). In the Ad libitum group, daily food intake increased, respectively, from the 2th to the 31st day from hatching from 63 5 to 8 1 g/day. There was then a spontaneous decrease to 6 9 g/day around the 6th day. The maximum and the minimum values of food intake were signi cantly different (Wilcoxon test, P<.1). In the Restricted group, there was no signi cant change of daily food intake during the studied period of

Growth and food intake in the European barn owl 141 Table 3. Body mass tted parameters obtained by the logistic equation for groups reared in different conditions. The equation was tted to each individual chick. Adult body mass is the body mass reached by the chicks of the experiment one year later. A is the asymptote, K the growth constant and ti the in ection point. Values are means S.D. Values of the same column with different labels are signi cantly different (Kruskal±Wallis test, P<.5; Dunn's test P<.5) Group n Adult body mass (g) A (g) K (day -1 ) ti (days) Control 2 317 39.6 11.5 a.177.3 a 17.18 1.35 a Ad libitum 5 337 39.6 35. a.174.16 a 17.67 1.37 a Restricted 5 317 338.7 8.26 b.215.53 a 16.26 2.3 a 1 5 1 5 Humerus Body mass Maximum Tarso metatarsus Fledging 92 46 73 36.5 317.7 23.9 g. There was no difference in edging body mass or its timing (MWU-test, P>.1 and Fig. 3b). The body mass loss between the asymptote and the edging mass was 72.9 17.8 g for the Ad libitum group which is signi cantly different from the 39.7 13.3 g of the Restricted group (MWU-test, P =.16). Feather growth parameters showed no signi cant difference between the two groups (MWU-test; A, P =.35; K, P =.176; and ti, P =.176). 1 5 1 5 1 5 Tibia Rectrix Remex 2 4 6 8 Age (days) 97 48.5 113 56.5 214 17 Fig. 2. Linear growth of various bones and feathers expressed as a percentage of asymptote or in mm for the Control group (logistic curves). growth (maximum vs. minimum food intake, Wilcoxon test, P =.9), despite a high value of 57 15 g/day on the 36th day. On a daily basis, the food intake of the Restricted group was signi cantly lower than in the Ad libitum group (MWU-test, P<.1). The food intake during the studied period from age 2 to 6 days in the Ad libitum group was on average 67 17 g/day, i.e. 17% higher than the Restricted group with a maximum difference of 28.5% at the Ad libitum peak of food intake. The major in uence of the restricted diet on body mass growth was the attening of the bell-shaped curve observed with the ad libitum diet (median asymptote signi cantly lower, Kruskal±Wallis test, P<.3; Dunn's test P<.5; no signi cant difference for the growth parameters K and ti, Kruskal±Wallis test, P=.51 and P =.651, respectively). Body mass at edging time was calculated using the tted equation for each chick. For the Restricted group, the value was 298.9 1.6 g and for the Ad libitum group DISCUSSION Compared to previous data (Wilson, Wilson & Durkin, 1987; Langford & Taylor, 1992; Taylor, 1994), the new information brought by the present study on barn owls was the measurement of the chicks' food intake during their growth. Our data enhance our knowledge on the signi cance of the bell-shaped curve encountered in many species of birds (Ricklefs, 1968; Olendorff, 1974; KorpimaÈki, 1981; Dijkstra, 1988; Lessells & Avery, 1989; Holt, Melvin & Steele, 1992; Hamer & Hill, 1993; Lacombe et al., 1994; Starck, KoÈnig & Gwinner, 1995; Kristan, GutieÂrrez & Franklin, 1996) by showing: I. That the decrease in body mass that occurs before edging is not due to a reduction of parental care but results from a spontaneous reduction of food intake. II. That a normal development occurs when the bellshaped growth curve, i.e. the overshoot, is attened by a reduction of food availability. Characteristic growth rate of the barn owl Field data are available on body mass and linear growth for various subspecies of barn owls (Pickwell, 1948; Ricklefs, 1968; SchoÈ nfeld & Girbig, 1975; Lenton, 1984; Wilson et al., 1987) and for the present subspecies, Tyto a. alba (GueÂrin, 1928; Bunn, Warburton & Wilson, 1982; Baudvin, 1986; Langford & Taylor, 1992; Taylor, 1994). There is no difference between our data and those found in other studies conducted on a larger number of birds for the same subspecies (for parameters calculated from GueÂrin's data 1928; for the asymptotic body mass, Baudvin, 1986; Langford & Taylor, 1992; Taylor, 1994; for the t 1±9 index, De Jond in Wijnandts, 1984). Therefore, we may conclude that our Ad libitum group exhibited the normal, characteristic, growth rate (Ricklefs, 1973; Lenton, 1984).

142 Our growth constant (K, body mass) for the European barn owl was.177 which is low if we consider the adult body size of the bird (Ricklefs, 1968). The European barn owl nests inside buildings which give a protection against predators and climatic variations. Furthermore, the female stays continuously with her brood during the rst 2 days of development. Ricklefs (1968) found that hole-nesting birds such as barn owls usually have a lower growth rate than open-nesting birds. This could explain the difference of growth rate encountered with other owl species of similar size (Asio otus, Wijnandts, 1984, or Asio ammeus, Holt et al., 1992). On the other hand, the characteristic growth rate certainly has an endogenous origin. The barn owl is a species originating from warm climates (Taylor, 1994). This could explain its low growth rate and longer breeding period in comparison with more northern species such as the short-eared or the long-eared owl. The Restricted group's diet simulates food intake of the youngest chick of a wild brood In barn owls, 2 days is a critical age in development. Chicks become thermally emancipated (Taylor, 1994) and begin to eat prey by themselves. Before 2 days, the chicks need female care to be fed with torn prey. The female starts to forage for long periods at night when her youngest chick is between 12 and 2 days old (Bunn et al., 1982; Taylor, 1994) and still unable to eat by itself. Consequently, when the female is foraging, the youngest chick may suffer food restriction because of its incapacity to eat the food available in the nest. The youngest chick can take its rst meal of the day only when the female stays in the nest at the end of the night to rest, a period when chicks seldom eat (Durant, pers. obs.). The youngest chick will then eat only during daytime (female resting period) like the chicks of our Restricted group. Thus, our Restricted group reproduces the situation experienced by the youngest chick in the wild. Since there was no apparent in uence of this hypocaloric diet on the timing of development in this experimental group, we may also assume that there is none for the youngest edgling of a wild brood in a year of normal prey availability. Food intake of captive birds is lower than in the eld For chicks fed ad libitum, we calculated the average daily food intake to be 67 17 g/day (Fig. 3a). This average value is overestimated because the calculation was made from the age of 2 days to edging, therefore not taking into account the period of lesser food intake the rst few days after hatching. From eld observations (Durant, Gendner & Handrich, 1996), we estimate the average number of prey brought to the nest per chick to be 2.9/day. With an average prey's mass of 23.6 g (Durant, pers. obs.), each chick could obtain 68.4 g a day, which is nearly the same as the average daily food intake we calculated. The main prey of the European barn owl during the brooding period in eastern France is the common vole Microtus arvalis (Muller, 1991). Using the average caloric values for laboratory mice of 7.35 kj/g (Handrich, Nicolas & Le Maho, 1993) and for the common vole of 7.74 kj/g (Wijnandts, 1984), we could calculate energy intake. The gross energy intake for wild chicks (529.4 kj/day) was then 7.4% greater than for captive chicks (493 kj/day). This difference might be due to an over-estimation of the number or size of prey actually delivered to a chick in the eld. Indeed, food storage is frequently observed in the nest (Baudvin, 198; Bunn et al., 1982; Taylor, 1994). Another explanation could be a problem of dehydration for the wild chicks because the temperature in certain nest sites (garrets, roofs or other protected locations) can be more than 1 8C higher than the outdoor ambient temperature (Durant, pers. obs.). Dehydration could increase the chicks' requirements, especially if they consume stored, and thus also partially dehydrated, prey. As a consequence, food ef ciency might be lower (rate of increase in body mass per unit food intake; BjoÈ rntorp & Yang, 1982; Brownell et al., 1986) in the eld than in captivity. Growth overshoot and spontaneous drop in food intake Our results on barn owls in captivity support the idea that the overshoot of body mass does not seem indispensable for the edgling (for petrels, Ricklefs & Schew, 1994; for kestrels, Lacombe et al., 1994). Why then do chicks undergo such a development? Lacombe et al. (1994) demonstrated that, despite the same developmental timing, chicks fed ad libitum and having a peak body mass were fatter than underfed chicks at edging. This observation is consistent with the energy insurance hypothesis (Ricklefs & Schew, 1994), which predicts that chicks accumulate fat as insurance against periods of poor feeding conditions. On the other hand, Lacombe et al. (1994) found a difference in fat reserves only for chicks reared with at least a 3% food restriction. The difference of overshoot between our two groups may not be only the result of the accumulation of energy reserves because we experimented with a food restriction of only 17%. At the body mass peak, the feather growth is at its maximum (ti in Table 2) and requires abundant protein and water (Murphy & King, 1991; Dietz, Daan & Masman, 1992). The reserves could be either proteins or water but the water loss when feather synthesis ends cannot explain the decrease in food intake we observed. Nevertheless, it could be advantageous to constitute reserves when possible instead of depending only on food supply when the need is the most important. When 5% of the feather growth is completed (ti), the chicks may start using the reserves to reduce the body mass for the edging. Since the feather development was normal in the Restricted group, the energy, protein and water inputs were suf cient. Thus, the reserves may be more a protection against a period of hypothetical food scarcity

Growth and food intake in the European barn owl 143 Food intake (g/day) 9 8 7 6 5 4 a α γ β γ Brood food intake (g) 32 3 28 26 24 22 Ad libitum feeding 3 b 2 18 Restricted feeding Body mass (g) 3 2 1 Body mass Remex 3 2 1 Remex length (mm) 3 4 5 6 7 First hatched chick's age (days) Fig. 4. Brood food requirement in relation to the age of the rst chick. Food requirement was calculated for a brood of four chicks with a hatching asynchrony of 2.5 days for both the Ad libitum group and the Restricted group. 1 2 3 4 5 6 7 Age (days) Fig. 3. Food intake and growth of the Ad libitum and Restricted groups. (a) Mean daily food intake based on a 4-day interval in relation to age (open bars for the Ad libitum group and cross-hatched bars for the Restricted group). The difference between extreme values (maximum and minimum) was tested. Values with different labels are signi cantly different (P<.5). The dashed curve corresponds to an extrapolation of daily food intake before food measurements began when the chicks were 2 days old. (b) Mean body mass and mean remex length in relation to age (solid curve for the Ad libitum group and dotted curve for the Restricted group). The open circles correspond to the mean logistic body mass curves. The signi cantly different part is delimited by the two vertical segments. (Ricklefs & Schew, 1994) than an accumulation of protein and/or water for the period of maximum growth rate. In times of food scarcity, the barn owl changes its diet (Taylor, 1994), which would change the intake of essential amino acids, e.g. cysteine (Murphy & King, 1991) for feather synthesis. In such a case, the limitation for feather growth would be the availability of the essential amino acids (protein). It would then be advantageous for a chick to have stored essential amino acids. Such a chick would experience normal growth at the same time that a chick with no overshoot would experience growth delay. The composition of the energy, water and protein reserves could be measured by body composition throughout the growing period. Apparently such a study has never been done. The amount of overshoot could also have an effect on the post- edging survival, a period when the young are forced to hunt by themselves. Since no difference was found for body mass and size at edging, the effect of the overshoot on subsequent survival could be due to only internal factors. Consequences of the overshoot value can be assessed only by body composition analysis. The present work demonstrates that the body mass decrease in ad libitum conditions, consecutive to the peak of body mass, is the consequence of a spontaneous decrease of food intake. What would be the advantage of this kind of growth for parental investment? We can calculate the brood requirement for a hypothetical brood fed ad libitum (Fig. 4; calculation made for a brood of four chicks with a hatching asynchrony of 2.5 days). The two-day period of maximum brood requirement (Fig. 4) in ad libitum conditions is obviously an important parameter of parental investment which cannot be modi ed. It depends only on the delay of the hatching asynchrony (in a synchronized brood, the duration for the maximum requirements would be 16 days). The short duration of the peak brood requirement is, in addition to the hatching asynchrony, a consequence of the spontaneous reduction of chick food intake when the asymptotic body mass is reached (Fig. 3a, b). For the brood, the energy that becomes available by the spontaneous decrease of food intake of a given chick could be transferred to its younger siblings at their maximum food requirement period. Resources would therefore not be wasted on the rst hatched chicks but used to improve the growth of younger siblings. Moreover, the maximum food requirement is dispatched over a longer period of time (Peak load hypothesis; Mock & Schwagmeyer, 199). As a consequence, brood reduction would not occur before a more severe food restriction. For the female, shortening the time of the peak brood requirements allows the transfer of a part of the food supply, i.e. the male investment, from the brood to the restoration of her body reserves in preparation for an eventual second brood. This would be an important bene t in a species like the barn owl where a second

144 brood could overlap with the end of the rst brood (Muller, 199). Thus, for the parents, an advantage of the bell-shaped growth curve that gives a greater exibility of chick requirements could be a decrease of the in uence of food availability on parental investment. If a comparison is made between the hypothetical brood fed ad libitum and another one fed with the restricted diet used in the present study (Fig. 4), we would expect a 2% decrease of the peak food intake (calculation made for a four-chick brood with a hatching asynchrony of 2.5 days). The parents could have raised at least one more chick with the restricted diet, but the youngest chick would experience a growth delay whenever there is a decrease of food supply (Lacombe et al., 1994). Because of this growth delay, even if the food shortage is of short duration, the youngest chick could suffer from starvation when the female is foraging. Moreover, the whole brood could suffer a growth delay if the food restriction lasts longer. Thus, this margin in food intake can be interpreted as a safety asset against small uctuations of food supply without brood reduction. CONCLUSION The food intake of barn owls in captivity undergoes a spontaneous decrease after the body mass has reached its maximum value. This spontaneous reduction of food requirement of the oldest chicks of a brood when the youngest are at their peak requirement could be an advantage for raising a brood successfully. A reduction of food availability does not seem to affect the development of barn owl chicks before edging except for a reduction of the maximum body mass. The body reserves that contribute to the overshoot of body mass seem to be a security against periods of food scarcity. As a consequence, a short period of food restriction would have no effect on the development (growth delay, brood reduction) since the reserves will cover the needs (energy, water and protein). Acknowledgements We thank Erkki KorpimaÈki, Henk Visser, Denver W. Holt, Yvon Le Maho, Charles A. Bost, Sylvie Massemin and members of the laboratory for helpful comments on various drafts of this paper. We also thank Christine DualeÂ, Fleur Semensatis and Guillaume Froget for their help with the English. REFERENCES Baudvin, H. (1978). Les causes d'eâchecs des nicheâes de chouettes effraies (Tyto alba). Le Jean le Blanc 17: 29±34. Baudvin, H. (198). Les surplus de proies au site de nid chez la Chouette Effraie (Tyto alba). Nos Oiseaux 35: 232±238. Baudvin, H. (1986). La Reproduction de la Chouette Effraie (Tyto alba). Le Jean le Blanc 25: 1±125. BjoÈrntorp, P. & Yang, M.-U. (1982). Refeeding after fasting in the rat: effects on body composition and food ef ciency. Am. J. Clin. Nutr. 36: 444±449. Brownell, K. D., Greenwood, M. C. R., Stellar, E. & Shrager, E. E. (1986). The effects of repeated cycles of weight loss and regain in rats. Physiol. Behav. 38: 459±464. Bunn, D. S., Warburton, A. B. & Wilson, R. D. S. (1982). The barn owl. Calton: T. & A. D. Poyser. Dietz, M. W., Daan, S. & Masman, D. (1992). Energy requirements for molt in the kestrel Falco tinnunculus. Physiol. Zool. 65: 1217±1235. Dijkstra, C. (1988). Reproductive tactics in the kestrel Falco tinnunculus. A study in evolutionary biology. Thesis, University of Groningen, Groningen. Duke, G. E., Evanson, O. A. & Jegers, A. (1976). Meal to pellet intervals in 14 species of captive raptors. Comp. Biochem. Physiol. 53A: 1±6. Durant, J., Gendner, J.-P. & Handrich, Y. (1996). A nest automatic weighing device to study the energetics of breeding barn owls (Tyto alba). International Conference on Raptors No. 2: 34. Edwards, T. C. (1987). Standard rate of metabolism in the common barn owl (Tyto alba). Wilson Bull. 99: 74±76. GueÂrin, G. (1928). La vie des chouettes. ReÂgime et croissance de l'effraye commune Tyto alba alba L. en VendeÂe. Paris: P. Lechevalier. Hamer, K. C. & Hill, J. K. (1993). Variation and regulation of meal size and feeding frequency in Cory's shearwater Calonectris diomedea. J. Anim. Ecol. 62: 441±45. Handrich, Y., Nicolas, L. & Le Maho, Y. (1993). Winter starvation in captive common barn-owls: bioenergetics during refeeding. Auk 11: 47±48. Holt, D. W., Melvin, M. & Steele, B. (1992). Nestling growth rates of short-eared owls. Wilson Bull. 14: 326±333. Johnson, W. D. (1974). The bioenergetics of the barn owl, Tyto alba. MA thesis, California State University, Long Beach. Kristan, D. M., GutieÂrrez, R. J. & Franklin, A. B. (1996). Adaptative signi cance of growth patterns in juvenile spotted owls. Can. J. Zool. 74: 1882±1886. KorpimaÈki, E. (1981). On the ecology and biology of Tengmalm's owl (Aegolius funereus) in Southern Ostrobothnia and SuomenselkaÈ, western Finland. Acta Univ. Ouluensis Ser. A Sci. Rerum Nat. Biol. 13: 1±84. Lack, D. (1968). Ecological adaptation for breeding in birds. London: Methuen. Lacombe, D., Bird, D. & Hibbard, K. A. (1994). In uence of reduced food availability on growth of captive American kestrels. Can. J. Zool. 72: 284±289. Langford, I. K. & Taylor, I. R. (1992). Rates of prey delivery to the nest and chick growth patterns of barn owls Tyto alba. In The ecology and conservation of European owls: 11±14. Galbraith, C. A., Taylor, I. R. & Percival, S. M. (Eds). Peterborough: U.K. Nat. Conserv. No. 5 Lenton, G. M. (1984). The feeding and breeding ecology of Barn owls Tyto alba in peninsular Malaysia. Ibis 126: 551±575. Lessells, C. M. & Avery, M. I. (1989). Hatching asynchrony in European bee-eaters Merops apiaster. J. Anim. Ecol. 58: 815± 835. Mock, D. W. & Schwagmeyer, P. L. (199). The peak load reduction hypothesis for avian hatching asynchrony. Evol. Ecol. 6: 249±26. Muller, Y. (199). Chevauchement des nidi cations successives chez la Chouette Effraie (Tyto alba). Alauda 58: 217±22. Muller, Y. (1991). La Chouette effraie (Tyto alba) dans la ReÂserve de la BiospheÁre des Vosges du Nord: uctuations des populations, reproduction et reâgime alimentaire. Ann. Sci ReÂs. Bios. Vosges du Nord 1: 91±16. Murphy, M. E. & King, J. R. (1991). Nutritional aspects of avian molt. Acta Congressus Internationalis Ornithologici 2: 2186± 2193.

Growth and food intake in the European barn owl 145 Olendorff, R. R. (1974). Some quantitative aspects of growth in three species of buteos. Condor 76: 466±468. Pickwell, G. (1948). Barn owl growth and behaviorisms. Auk 65: 339±373. Ricklefs, R. E. (1967). A graphical method of tting equations to growth curves. Ecology 48: 978±983. Ricklefs, R. E. (1968). Patterns of growth in birds. Ibis 11: 419± 451. Ricklefs, R. E. (1973). Patterns of growth in birds 2. Growth rate and mode of development. Ibis 115: 177±21. Ricklefs, R. E. & Schew, W. A. (1994). Foraging stochasticity and lipid accumulation by nestling petrels. Funct. Ecol. 8: 159± 17. SchoÈnfeld, M. & Girbig, G. (1975). BeitraÈge zur Brutbiologie der Schleiereule, Tyto alba, unter besonderer BeruÈcksichtigung der AbhaÈngigkeit von der Feldmausdichte. Hercynia 12: 257±319. Starck, J. M., KoÈnig, S. & Gwinner, E. (1995). Growth of stonechats Saxicola torquata from Africa and Europe: an analysis of genetic and environmental components. Ibis 137: 519±531. Taylor, I. (1994). Barn owls: predator±prey relationships and conservation. Cambridge: Cambridge University Press. Wijnandts, H. (1984). Ecological energetics of the long-eared owl (Asio otus). Ardea 72: 1±92. Wilson, R. T., Wilson, M. P. & Durkin, J. W. (1986). Breeding biology of the barn owl (Tyto alba) in central Mali. Ibis 128: 81±9. Wilson, R. T., Wilson, M. P. & Durkin, J. W. (1987). Growth of nestling barn owls Tyto alba in central Mali. Ibis 129: 35±318.