Ectoparasitism in marsh tits: costs and functional explanations

Behavioral Ecology Vol. 14 No. 2: 175 181 Ectoparasitism in marsh tits: costs and functional explanations Jan-Åke Nilsson Department of Animal Ecology, University of Lund, S-223 62 Lund, Sweden Among hole-nesting birds, the blood-sucking hen flea is a common parasite affecting both nestlings and parents. By adding fleas to marsh tit (Parus palustris) nests, I aimed to investigate the effect of fleas on nestling growth rate and parental effort as well as evaluating a potential mechanism by which fleas affect nestlings (i.e., resting metabolic rate; RMR). Nestlings from flea-infested broods were lighter than nestlings from control broods. This reduced growth rate was evident as soon as 3 days after adding extra fleas to the nest, but the reduction did not increase after this initial drop in mass. Parents did not alter their feeding frequency in response to the manipulation; thus the small size of nestlings in manipulated nests seems to be directly caused by the fleas. Massspecific RMR was significantly higher in nestlings from flea-infested nests compared to nestlings from control nests. I used the results to evaluate the suggested mechanisms for parasite-related decrease in host growth rate. The increase in RMR and the very rapid reduction in nestling growth rate after experimental addition of fleas can be explained by an immune reaction, mainly by the innate immune system, to substances in the saliva of the fleas. Key words: ectoparasitism, fleas, innate immune system, marsh tits, metabolic rate, nestling condition, parental effort, Parus palustris. [Behav Ecol 14:175 181 (2003)] Parasites derive their resources from the bodies of their hosts, thereby potentially affecting these hosts negatively. A host parasite relation that has received considerable interest during the last decade is the one between hematophagous ectoparasites and nesting birds. This system has several advantages, including its suitability for experimental manipulation and the possibility to directly assess the effect of the parasite on survival and reproduction of its host. The ability of parasites to live for a long time in the nest material (Rothschild and Clay, 1952), thereby giving them opportunity to be transmitted horizontally (Heeb et al., 1996), indicate that the effect on host fitness might be severe (Clayton and Tompkins, 1994). In this host parasite system, two types of hosts are available to the parasites: parents and nestlings. Because nestlings are bound to the nest, they will be especially vulnerable to ectoparasites that live in the nest material, such as the hen flea (Ceratophyllus gallinae). Consequences of experimentally manipulating the abundance of ectoparasites on nestlings include reduced growth rate and survival (Allander, 1998; Dufva and Allander, 1996; de Lope et al., 1993; Merino and Potti, 1996; Møller, 1990, 1997; Møller et al., 1990; Richner et al., 1993). But parents, especially females, also spend some time in the nest, during brooding and feeding of the young, for example. These behaviors put them at risk of being exposed to the ectoparasites, potentially leading to a decrease in survival probability (Brown et al., 1995). Survival of parents may also be affected by ectoparasites in an indirect way. Lifehistory theory will, under certain circumstances, predict an increase in parental effort to meet the increased demands of parasitized nestlings. Blue tit (Parus caeruleus) parents have been shown to increase feeding rate to heavily infested broods (Tripet and Richner, 1997), leading to reduced future reproductive success (Richner and Tripet, 1999) and leading pied flycatcher (Ficedula hypoleuca) females to increase energy Address correspondence to J.-Å. Nilsson. E-mail: jan-ake.nilsson@ zooekol.lu.se. Received 26 October 2001; revised 24 April 2002; accepted 17 June 2002. Ó 2003 International Society for Behavioral Ecology expenditure at least up to a threshold of infestation (Merino et al., 1998). Thus, hematophagous ectoparasites exert fitness costs to both parents and nestlings. These costs seem to be dependent on the amount of resources available to nestlings and parents, as the negative effects of ectoparasites increase during harsh environmental conditions (Dufva and Allander, 1996; de Lope et al., 1993). Several mechanisms have been proposed to explain this increased resource requirement of parasitized birds, but empirical support for any of them is generally lacking. The increased demand for energy and nutrients suggested to be caused by ectoparasites include compensation for the blood removed from hosts, fighting off secondary infections either transmitted by the ectoparasite or received due to the open wounds caused by ectoparasites, mounting an immune response to substances in the saliva of ectoparasites, and performing costly behaviors directed to reduce the impact of parasites. Although an increase in energy demands of parasitized hosts is hypothesized, few studies of actual energy turnover rates of hosts have been performed. The only study investigating the relationship between degree of hematophagous ectoparasitism and rate of energy turnover, to my knowledge, comes from the system with the house martin bug (Oeciacus hirundinis) as an ectoparasite on the house martin (Delichon urbica; Møller et al., 1994). The massspecific daily energy expenditure of house martin nestlings increased with the level of parasite infestation. The increased energy expenditure was suggested to be mediated through increased rates of begging in parasitized broods (Møller et al., 1994) because begging rate has been shown to increase as a result of heavy infestation (Christe et al., 1996a). The hematophagous hen flea is a common ectoparasite on adults and nestlings of cavity-breeding passerines such as tits. Adult fleas are dependent on blood from their hosts for survival and reproduction (Rothschild and Clay, 1952). Previous studies on the effect of fleas on nestlings of tits have produced contradictory results. Great tit (Parus major) nestlings have repeatedly been shown to suffer severely from the action of fleas (Allander, 1998; Dufva and Allander, 1996; Heeb et al., 1999; Richner et al., 1993), whereas blue tit nestlings seem to remain largely unaffected by fleas (Tripet and Richner, 1997). Here I present data on a third tit species,

176 Behavioral Ecology Vol. 14 No. 2 the marsh tit (Parus palustris), which is the same size as the blue tit. The aim of this study was to experimentally investigate how fleas affect parent and nestling marsh tits. Specifically, I wanted to (1) study the effect of the parasite on nestling growth rate, (2) look for adjustments in parental effort to alleviate the effects on nestlings, (3) investigate one potential mechanism by which fleas affect nestlings (i.e., relating resting metabolic rate [RMR] of nestlings to degree of infestation). METHODS General I studied parental breeding effort and nestling growth in marsh tits during the breeding season of 1997 in a 64-km 2 large study area, 20 km east of Lund, southern Sweden. The study area consists of small deciduous forests and groves interrupted by permanent pastures and agricultural fields. Forested areas were provided with identical nest-boxes and all broods investigated in this study originated from these nest-boxes. I checked nest-boxes at least once a week to be able to calculate the day of the first egg (assuming that one egg was produced each day, which is the general rule; own observation) and to determine clutch size (mean: 9.6 eggs; range: 7 11). Toward the end of the incubation period, all boxes were checked daily to determine hatching date (median: 21 May; range: 12 28 May). All age references of nestlings given below are in relation to hatching day, which is denoted day 0. When nestlings were 3, 6, 9, 12, and 15 days old, I weighed all individuals within a brood to the nearest 0.1 g. The weighings on day 3 were performed before any broods had been manipulated and are referred to as mean initial mass. On nestling days 9, 12, and 15, I also measured wing length to the nearest 0.5 mm. Caution was taken to measure each brood at about the same hour of the day during the entire study. Nestlings were banded for individual recognition when they were 6 days old. When nestlings were 15 days old, most parents were captured and measured in the same way as nestlings. However, one female and two males eluded capture on this day and were instead measured at day 16. Toward the end of the nestling period, I checked the boxes daily to determine fledging date. Manipulation Broods were randomly divided into two groups, with the only restriction that the groups should contain broods from the entire span of hatching dates. One of these groups was left unmanipulated and served as control broods. Broods in the other group, the experimental group, received 30 adult fleas when nestlings were 3 and 9 days old, thus, in total 60 extra fleas. Fleas were added to the nest-box after completion of the weighings for that day. I extracted fleas for experimental manipulation from old nests during spring and stored them in a refrigerator until used. In total, 11 nests were experimentally flea infested and 9 nests served as controls. Between day 10 and 12, one control nest was depredated, thus reducing sample size for the later measurements. On the day of nest leaving, I collected all nests and placed them in a plastic funnel with a plexiglass cover to prevent fleas from escaping. Each nest was kept in its funnel for 3 days with continuous illumination from above. Fleas were collected in containers beneath the funnel, and the counts served as an estimate of relative flea abundance. Experimentally infested nests contained a median number of 19 fleas (range: 13 35 fleas), whereas control nests only contained 6 fleas (range: 0 26 fleas). Thus, my manipulation was successful in creating infested nests with significantly more fleas than in control nests (Mann-Whitney U test: U 5 78.5, N 5 19, p 5.027), even 11 12 days after the last experimental addition of fleas. Feeding frequency On nestling day 9, I captured the feeding male and female to attach a small transponder (length: 11.5 mm; diam: 2.12 mm; Trovan, ID 100) to their color rings. The mass of a transponder is 0.09 g, thus increasing the mass of an average marsh tit by only 0.8%. Each transponder had an individual code, which was registered with the help of a circular antenna applied around the entrance hole inside the nest-box. The identification code, together with the time of registration, was stored on a data logger (Trovan, LID 604). All feeding visits of female and male parents were registered in this way during 2 full days, starting when nestlings were 10 days old and ending on day 12. Feeding during this period is taken to represent relative parental effort of the females. RMR measurements The resting metabolic rate (RMR) of individual nestlings was measured in an open-flow respirometer during night. Restriction of measurement units in the respirometer, four channels are available, resulted in the failure to fit in young from one control brood. Depending on the number of free channels in the respirometer, I measured RMR on one (n 5 9) or two (n 5 10) randomly selected nestlings in the remaining broods. I entered only one nestling in each brood into the analyses, and in the cases where two nestlings from the same brood were measured, I included the one with a mass closest to the mean mass of all measured nestlings. The reason for this was to reduce the mass differences between nestlings from the two brood categories as much as possible. Entering the mean RMR of the two nestlings did not change any of the results. Late in the evening, when nestlings were 15 (n 5 14) or 16 (n 5 5) days old, they were individually placed in a sealed respirometer chamber (1.6 l) and placed in the darkness of a temperature-controlled cabinet (BK600, Heraeus, Hanau, Germany) at 258C, i.e., within their thermoneutral zone (Gavrilov and Dolnik, 1985). After some minutes of movement, as judged from the oxygen consumption readings, the birds settled and produced a very smooth and even oxygen consumption level. Measurements ended in early morning, and nestlings were immediately returned to their nest-box. Air was sucked through the respirometer chamber containing the bird at a rate of 10.0 l/h (flow controls: F-111C, Bronkhorst HI-TEC, Ruurlo, the Netherlands) and dried with the help of silica gel. The oxygen concentration of this air was analyzed to the nearest 0.01% (Xentra 4100 gas purity analyzer, Servomex, UK) and compared to the oxygen concentration in the reference air for calculations of oxygen consumption (Klaasen et al., 1997). Oxygen concentration was automatically recorded on a Grant Squirrel data logger (model 1202) every minute. Reference air was measured for 15 min during each 90-min period throughout the measurement session. The value of oxygen consumption (ml O 2 /min) used in the analyses was taken as the single lowest value of running 10-min averages during a measurement session. Oxygen consumption was very stable during the latter part of the night, resulting in long periods with this lowest value. I converted oxygen consumption to metabolic rate (kj/h) by assuming an energetic equivalence of 19.8 kj/l O 2. I tested the residuals for normality (Lilliefors test) in all statistical tests and used nonparametric statistics if the residuals were not normally distributed.

Nilsson Costs of ectoparasitism 177 Table 1 Summary of analyses of covariance (ANCOVA) with mean mass of nestlings (N 5 19 broods) at different ages as dependent variable and the potential explanatory variables hatching date, number of nestlings, initial mass, age of male, age of female, and experimental category (5 flea-infested or control nest) Age Variable Part. corr. F p R 2 Figure 1 Mean (6 SE) mass (g) of nestlings in unmanipulated, control nests (open circles; broken line) and in flea-infested nests (filled circles; solid line). Mass was measured at nestling ages of 3, 6, 9, 12, and 15 days. Differences between nest categories tested with repeatedmeasures ANOVA: F 1,16 5 20.87, p,.001. Interaction between growth rate and experiment: F 5 3.88, p 5.007. RESULTS Clutches giving rise to control and flea-infested broods did not differ in hatching date (t test: t 18 5 0.1, p 5.9) or in size (t 18 5 0.7, p 5.5). Neither did the age of males or females differ between the two categories (chi-square test: p..1 in both males and females). Furthermore, masses of 3-day-old nestlings (i.e., before the first flea-infestation) were similar in the two experimental groups (controls: mean 5 3.0 g; SD 5 0.51; flea-infested: mean 5 2.9 g; SD 5 0.59; t test: t 17 5 0.46, p 5.6). Day 6 Date.31 5.71.030.76 Initial mass.54 17.3.001 Experiment 11.7.004 Day 9 Initial mass.49 8.11.012.61 Experiment 14.1.002 Day 12 Experiment 8.30.010.33 Day 15 No. nestlings 2.39 5.76.029.59 Experiment 19.0.001 The final model was obtained by stepwise, backward elimination of nonsignificant variables. The partial correlation coefficients (part. corr.) are presented for linear variables. Nestling growth and survival Nestlings survived well in both brood categories. Only three nestlings died; one in a control and two in flea-infested broods. These young died early during the nestling period (between day 3 and 9) and were always the smallest in their respective broods. Already at day 6, 3 days after the first addition of fleas to the manipulated nests, nestlings in flea-infested nests had a significantly lower mass than nestlings in control broods (Figure 1, Table 1). After this age, the difference in mean mass remained at the same level, as evidenced by a nonsignificant interaction between growth rate and experiment (repeated-measures ANOVA [days 6 15]: F 1,17 5 15.08, p 5.001, interaction: F 5 0.23, p 5.88). Thus, the relative difference was largest at day 6 when the mean mass of nestlings from flea-infested broods was 10.1% lower than in control broods. Beside the experimental effect, other variables explained some of the variation in mean mass at different ages (Table 1). At young ages, initial mass still had an effect on mean mass, but this influence disappeared after day 9. Additionally, the number of young in the brood had a negative effect on mean mass late in the nestling period. However, at most ages the experimental addition of fleas had the greatest effect on nestling mass (Table 1). Variation in mass within a brood, measured as standard deviation, did not at any age differ between flea-infested and control broods (t test: p..4 in all cases). Wing length of the nestlings were also significantly affected by the experiment, mean wing length of nestlings in fleainfested nests being shorter than in control nests (Figure 2, Table 2). The difference between the brood categories remained the same during the period of measurements (Figure 2). Thus, the relative difference in wing length between nestlings from flea-infested and control broods was largest at the youngest age (9.7 %), at day 9. The effect of the flea experiment was not as important, compared to other variables, for explaining the variation in mean wing length as it was for explaining variation in mean mass (Table 2). Also, wing length was affected by initial mass in the early part of the nestling period and was also affected by hatching date late in the nestling period. Variation in wing length within a brood was not related to flea manipulation at any of the ages (t test; p..3). The nestlings stayed on average 20 days in the nest (range: 19 22 days). Nestlings in flea-infested broods tended to have a slightly longer nestling period than those in control broods (t test: t 15 5 1.86, p 5.08). Since the experiment also affected other variables during breeding, I included date, number of nestlings, mean mass at day 15, mean wing length at day 15, feeding frequency of male and female, and experimental category as independent variables in a stepwise, backward multiple regression analysis with the length of the nestling period as the dependent variable. Two variables remained in the final model; date (partial correlation 5.56; p 5.028) and wing length at day 15 (partial correlation 52.84; p 5.002). Thus, the tendency for experimental category to influence the length of the nestling period was probably indirect through the effect of the experiment on mean wing length because having a short wing on day 15 significantly increased the time before fledging. Parental effort Parents did not significantly change their feeding frequency as a result of the experiment. Both males and females had more or less the same feeding frequency if they tended a flea-infested brood (mean 6 SE, males: 14.2 6 1.68, females: 18.3 6 1.71 feeds/h) as if they tended a control brood (males: 13.4 6 0.84, females: 15.4 6 1.46 feeds/h; t test: p..25 for both sexes). Neither did the number of feedings to individual young differ significantly between nestlings in the two brood categories (t test: t 16 5 1.30, p 5.21). Another way of looking for effects of parental work load would be through masses of the parents toward the end of the

178 Behavioral Ecology Vol. 14 No. 2 Table 2 Summary of ANCOVA with mean wing length of nestlings (N 5 19 broods) at different ages as dependent variable and the potential explanatory variables hatching date, number of nestlings, initial mass, age of male, age of female, and experimental category (5 flea-infested or control nest) Age Variable Part. corr. F p R 2 Day 9 Initial mass.62 15.4.001.61 Experiment 6.75.019 Day 12 Initial mass.43 5.50.033.53 Experiment 7.59.015 Day 15 Date.50 7.00.018.44 Experiment 4.98.040 Figure 2 Mean (6 SE) wing length (mm) of nestlings in unmanipulated, control nests (open circles; broken line) and in flea-infested nests (filled circles; solid line). Wing length measured at nestling ages of 9, 12, and 15 days. Differences between nest categories tested with repeated-measures ANOVA: F 1,17 5 6.79, p 5.018. Interaction between growth rate and experiment: F 5 1.04, p 5.36. nestling period. Females tending flea-infested broods weighed significantly less than those tending control broods (Figure 3). Males feeding in flea-infested broods also had masses that were lower than for males feeding in control broods, but this difference was not significant (Figure 3). To see which factors influenced parental mass beside the brood categories, I used date, number of nestlings, age, and feeding frequency together with brood category as independent variables in analyses of covariance (ANCOVA) for the sexes separately. A final model was obtained by stepwise, backward elimination of nonsignificant factors. Mass of males was negatively related to their feeding frequency (partial correlation coefficient: 2.58, F 5 9.53, p 5.009) and tended to be lower for males tending flea-infested broods (F 5 4.00, p 5.067). These two variables explained 55% of the variation in male mass. Only the brood category, flea-infested or control broods, explained a significant part of the variation in female mass (R 2 5.26, p 5.026; Table 3, Figure 3). Parental survival Overall survival to the next breeding season was similar for the two sexes: 72% for males and 68% for females. To see if feeding in control or flea-infested broods and parental mass could predict survival, these two factors were entered as independent variables in a logistic regression analysis with survival as the binary dependent variable. The discrepancy between the model and the data is reported as the deviance, which is distributed asymptotically as v 2. Variation in male survival could not be significantly predicted by either of the two factors (p..3). In females, flea manipulation could not explain any of the variation in survival; however, the probability of survival among females increased significantly with an increase in mass (treatment: deviance 5 0.8, df 5 1, p..1, mass: deviance 5 4.5, df 5 1, p 5.035). Nestling RMR Nestlings from flea-infested broods had a 10.3% higher RMR during night (mean 5 1.18 kj/h, SE 5 0.06, n 5 11) than had those from control broods (mean 5 1.07 kj/h, SE 5 0.03, n 5 8), but the difference was not statistically significant The final model was obtained by stepwise, backward elimination of nonsignificant variables. The partial correlation coefficients (part. corr.) are presented for linear variables. (t test: t 17 5 1.47, p 5.16). As in the total data set, nestlings from flea-infested broods were significantly lighter than those from control broods (t 17 5 2.52, p 5.022). Because mass influences metabolic rate, I calculated a mass-specific metabolic rate for each nestling. RMR per gram was significantly higher for nestlings from flea-infested broods than for those raised in control broods (Figure 4). Furthermore, mass-specific RMR was positively related to the relative abundance of fleas (p 5.027; Figure 5), indicating a causal relation between flea abundance and nestling metabolic rate. DISCUSSION Fitness costs of parasitism Nestlings from flea-infested broods were lighter and had a somewhat shorter wing than nestlings from control broods. Because parents did not reduce the number of feedings to flea-infested broods, the small size seems to be directly caused by the fleas. The cost of parasitism for the nestlings will be in the form of reduced survival either in the nest, although this did not happen during the conditions prevailing in the present study, or after nest-leaving because fledging mass is an important predictor of survival to the next breeding season (Tinbergen and Boerlijst, 1990). Furthermore, the shorter wing of parasitized young resulted in longer time in the nest with the increased potential risk of being taken by nest predators. Parents, especially females, provisioning flea-infested broods, were lighter at the end of the nestling period. Furthermore, females with low body mass had a reduced survival probability to the next breeding season. Fleas may affect survival of parents in two different ways, either directly through the actions of the parasite or indirectly through effects on parental effort. In general, in comparison with their young, parents are much less accessible to fleas due to the short duration of their visits to the nest. However, females are not only feeding their young. At least during the first week, females brood their young intermittently during daytime (Perrins, 1979) as well as spending the night in the nest for another week (Nilsson, personal observation). This could explain the sex differences in the consequences of tending flea-infested broods, although the treatment had no direct effect on female survival. Parental effort may be indirectly affected by parasitism in two different ways. Parents may increase their survival by reducing breeding effort to heavily flea-infested broods because the action of the ectoparasites

Nilsson Costs of ectoparasitism 179 Table 3 Summary of ANCOVA with mean mass of males and females as dependent variable and the potential explanatory variables hatching date, number of nestlings, age of male and female, feeding frequency of male and female, and experimental category (5 flea-infested or control nest) Sex Variable Part. corr. F p R 2 Male Feeding frequency 2.58 9.53.009.55 Experiment 4.00.067 Female Experiment 5.99.026.26 The final model was obtained by stepwise, backward elimination of nonsignificant variables. The partial correlation coefficients (part. corr.) are presented for linear variables. Figure 3 Mean (6 SE) mass at day 15 posthatching of male and female parents feeding nestlings in unmanipulated control (CON) nests and in flea-infested, experimental (EXP) nests. Differences between nest categories tested with t test: males: t 15 5 1.69, p 5.11; females: t 17 5 2.45, p 5.026. diminishes the value of current reproduction (Møller, 1994; Møller et al., 1994). Alternatively, they may increase parental effort to compensate for the adverse effects inflicted by the parasites on their young (Christe et al., 1996a; Tripet and Richner, 1997). In this study, parent marsh tits did not alter a major component of parental effort (i.e., their feeding frequency), in response to the manipulated abundance of fleas. However, parents may increase other aspects of reproductive effort, such as increased search effort when foraging, in response to increased diet selectivity of nestlings in flea-infested nests (see below). Furthermore, ectoparasite infestation may call for increased effort directed to nest sanitation (Christe et al., 1996b; Hurtrez-Boussès et al., 2000). Female great tits increased time devoted to nest sanitation at the expense of time for sleeping, thus increasing parental effort with potential consequences for survival (Christe et al., 1996b). Mechanisms responsible for the cost of parasitism The reduced growth rate of nestlings from flea-infested broods at the same time that food provisioning rate per nestling was the same indicates an increase in energy expenditure of these young. In line with this, adult great tits injected with a novel antigen (sheep red blood cells) increased their mass-specific RMR compared to sham-injected birds (Ots et al., 2001). The few energy measurements of hosts parasitized with natural parasites performed so far have yielded varying results depending on the host parasite system. Intestinal parasites, disrupting the digestive abilities of the host, resulted in reduced metabolic rates, at least when the animals were stressed (Connors and Nickol, 1991; Munger and Karasov, 1989). A pathogenic challenge with nematodes increased the metabolic rate of red grouse (Lagopus lagopus scoticus; Delahay et al., 1995). Featherfeeding ectoparasites have been shown to increase metabolic rate due to increased thermoregulatory costs (Booth et al., 1993). In the only study on host energetics involving a hematophagous ectoparasite, Møller et al. (1994) found that the mass-specific daily energy expenditure of house martin nestlings increased in response to parasitism by the house martin bug. The increased energy expenditure was suggested to depend in part on more intense begging behavior by highly infested nestlings. Increased begging rates of great tit nestlings have later been found to be a consequence of high flea abundances (Christe et al., 1996a). In this study, mass-specific RMR was 17% higher for fleainfested nestlings compared to control nestlings. This precludes an effect of altered nestling behavior in response to parasitism because RMR was measured during night when the nestlings are inactive in the respiratory chamber. Thus, the slower growth rate of nestlings in flea-infested broods seems depend on the action of the fleas, at least partly through an increase in the resting metabolism. One could argue that the difference in mass between control nestlings and nestlings from flea-infested broods is mostly due to different levels of metabolically inactive adipose tissue. If so, a difference in mass-specific RMR would merely reflect this structural difference. In this case, an increase in mass for a control nestling would result in a smaller increase in RMR than would be the case for a nestling in a flea-infested brood for the same increase in mass. Thus, the slope of the relation between RMR and mass would be more shallow for control nestlings than for flea-manipulated nestlings, resulting in a significant interaction between treatment and mass in relation to RMR. This was not the case (treatment 3 mass: p..8), indicating that nestlings from different categories had similar body composition. This conclusion was also reached by Burness et al. (2000) after measuring metabolic rate, mass of various organs, and lipid mass in nestlings from reduced and enlarged tree swallow (Tachycineta bicolor) broods. In that study, although the rearing environment differed, resulting in Figure 4 Mean (6 SE) mass-specific resting metabolic rate (RMR) of nestlings raised in flea-infested (EXP) and control (CON) broods. Differences between brood categories tested with t test: t 17 5 2.25, p 5.038.

180 Behavioral Ecology Vol. 14 No. 2 Figure 5 Mass-specific resting metabolic rate of marsh tit nestlings in relation to an index of flea abundance in the nest after fledging. Tested with regression analysis: r 5.51, n 5 19, R 2 5.26, p 5.027. a significant mass difference between nestlings in reduced and enlarged broods, the physiological development was relatively invariant. Furthermore, the mass-specific metabolic rate was close to significantly higher for the heavier nestlings in reduced broods compared to nestlings from enlarged broods (Burness et al., 2000). Which mechanisms can account for both the increase in nestling RMR and for the rapid reduction in nestling growth rate following experimental addition of fleas? Suggested possible mechanisms include forming of blood cells to compensate for losses to the fleas and mounting costly immune responses to substances in the saliva of fleas or to secondary infections. Compensation of blood loss Blood drainage by ectoparasites may be expected to cause anemia in nestlings. Low levels of hematocrit, indicative of anemia, have been found among nestlings in heavily infested nests (e.g., Chapman and George, 1991; Hurtrez-Boussès et al., 1997; Potti et al., 1999; Richner et al., 1993), although in other studies hematocrit has remained unaffected by parasitism (e.g., Johnson and Albrecht, 1993; Johnson et al., 1991; Saino et al., 1998). However, this effect should increase in importance with time of exposure to ectoparasites as their consumption of blood cumulates over time. The amount of blood removed by a flea has been estimated to 0.1 ll per day (Busvine, 1976). This can be compared with the standard blood sample of 20 ll (e.g., Richner et al., 1993) for measuring hematocrit, thus equivalent to the action of 200 fleas. The blood drainage action of the added fleas distributed among, on average, nine nestlings in my manipulated nests thus does not seem to have the potential to explain the rapid (within 3 days) reduction in nestling mass (see de Lope et al., 1998 for a similar conclusion). Repeated challenges of the immune system Even if no pathogens are transmitted by fleas, compounds in their saliva will be recognized as antigens, thus triggering an immune response (Jones, 1996; Nelson, 1987), as has been shown in some studies (de Lope et al, 1998; Szép and Møller, 1999). In nestlings, an immune response is primarily composed of elements from the innate immune system. This is because the production of T and B lymphocytes, making up the adaptive immune system, does not reach adult efficiency until several weeks after hatching (Apanius, 1998; Klasing and Leshchinsky, 1999). Thus, nestlings have to depend on their innate immune system to take care of antigens. This line of defense is primarily composed of phagocytic cells (e.g., monocytes and macrophages). When macrophages/monocytes are stimulated, they produce proinflammatory monokines (hormonelike peptides; e.g., interleukin-1), which function as signals from the immune system to other parts of the body including the brain, resulting in behavioral, cellular, and metabolic changes. Among the changes induced by these monokines are reduced gross food intake, fever, production of acute-phase proteins, and a diet shift away from proteins toward carbohydrates (Klasing and Johnstone, 1991; Maier and Watkins, 1999). Anorexia is considered to be an adaptive response to immune system challenges. It is suggested to result from increased diet selectivity (Kyriazakis et al., 1998), at least in part due to the monokine-induced diet alteration toward carbohydrates (Aubert et al., 1995). In a study by Klasing et al. (1987), stimulation of the innate immune system in chickens as well as direct administration of interleukin-1 resulted in reduced growth rates, which could be explained partly by a decreased food intake and partly by an increased core body temperature (i.e., fever). Fever is an adaptive increase in the set point for body temperature, induced by signals mediated by interleukin-1 to the brain (Maier and Watkins, 1999). Fever is also energy intensive, entailing an increased metabolic cost. An increase in core body temperature of 18C increases the metabolic cost by 10 15% (Beisel, 1977; Blaxter, 1989; Maier and Watkins, 1999). Thus, repeated stimulation of phagocytes by nonpathogenic immunogens commonly induces anorexia (e.g., Klasing et al., 1987, 1991; Kyriazakis et al., 1998) and fever (Klasing and Leshchinsky, 1999). The innate defense system is activated within 1 2 h of an immunological challenge (Maier and Watkins, 1999), therefore potentially explaining the rapid reduction in growth rate found in this study, both as a consequence of the cost of sustaining a higher body temperature and of a reduced food intake. Before reaching definite conclusions about potential functional explanations underlying the detrimental effects of ectoparasites, data on nestling rate of food intake and diet selectivity in connection with the manipulation as well as nestling body temperature are needed. I am grateful to M. Stjernman for field assistance and to D. Hasselquist, L. Råberg, and M. Stjernman for valuable comments on a previous version of the manuscript. This study was approved by the Malmö/Lund Ethical Committee and supported by grants from the Swedish Natural Science Research Council. REFERENCES Allander K, 1998. The effects of an ectoparasite on reproductive success in the great tit: a 3-year experimental study. Can J Zool 76:19 25. Apanius V, 1998. Ontogeny of immune function. In: Avian growth and development. (Starck JM, Ricklefs RE, eds). Oxford: Oxford University Press; 203 222. Aubert A, Goodall G, Dantzer R, 1995. Compared effects of cold ambient temperature and cytokines on macronutrient intake in rats. Physiol Behav 57:869 873. Beisel WR, 1977. Metabolic and nutritional consequences of infection. In: Advances in nutritional research, vol. 1 (Draper HH, ed). New York: Plenum Press; 125 143. Blaxter K, 1989. Energy metabolism in animals and man. Cambridge: Cambridge University Press. Booth DT, Clayton DH, Block BA, 1993. Experimental demonstration of the energetic cost of parasitism in free-ranging hosts. Proc R Soc Lond B 253:125 129.

Nilsson Costs of ectoparasitism 181 Brown CR, Brown MB, Rannala B, 1995. Ectoparasites reduce longterm survival of their avian host. Proc R Soc Lond B 262:313 319. Burness GP, McClelland GB, Wardrop SL, Hochachka PW, 2000. Effect of brood size manipulation on offspring physiology: an experiment with passerine birds. J Exp Biol 203:3513 3520. Busvine JR, 1976. Insects, hygiene and history. London: Athlone Press. Chapman BR, George JE, 1991. The effects of ectoparasites on cliff swallow growth and survival. In: Bird-parasite interactions ecology, evolution, and behaviour. (Loye JE, Zuk M, eds). Oxford: Oxford University Press; 69 92. Christe P, Richner H, Oppliger A, 1996a. Begging, food provisioning, and nestling competition in great tit broods infested with ectoparasites. Behav Ecol 7:127 131. Christe P, Richner H, Oppliger A, 1996b. Of great tits and fleas: sleep baby sleep... Anim Behav 52:1087 1092. Clayton DH, Tompkins DM, 1994. Ectoparasite virulence is linked to mode of transmission. Proc R Soc Lond B 256:211 217. Connors VA, Nickol BB, 1991. Effects of Plagiorhynchus cylindraceus (Acanthocephala) on energy metabolism of adult starlings, Sturnus vulgaris. Parasitology 103:395 402. Delahay RJ, Speakman JR, Moss R, 1995. The energetic consequences of parasitism effects of a developing infection of Trichostrongylus tenuis (Nematoda) on red grouse (Lagopus lagopus scoticus) energybalance, body-weight and condition. Parasitology 110:473 482. de Lope F, González G, Pérez JJ, Møller AP, 1993. Increased detrimental effects of ectoparasites on their bird hosts during adverse environmental conditions. Oecologia 95:234 240. de Lope F, Møller AP, Cruz C de la, 1998. Parasitism, immune response and reproductive success in the house martin Delichon urbica. Oecologia 114:188 193. Dufva R, Allander K, 1996. Variable effects of the hen flea Ceratophyllus gallinae on the breeding success of the great tit Parus major in relation to weather conditions. Ibis 138:772 777. Gavrilov VM, Dolnik VR, 1985. Basal metabolic rate, thermoregulation and existence energy in birds: world data. Acta Int Ornithol Congr 18:421 466. Heeb P, Werner I, Kölliker M, Richner H, 1998. Benefits of induced host responses against an ectoparasite. Proc R Soc Lond B 265:51 56. Heeb P, Werner I, Mateman AC, Kölliker M, Brinkhof MWG, Lessells CM, Richner H, 1999. Ectoparasite infestation and sex-biased local recruitment of hosts. Nature (Lond) 400:63 65. Heeb P, Werner I, Richner H, Kölliker M, 1996. Horizontal transmission and reproductive rates of hen fleas in great tit nests. J Anim Ecol 65:474 484. Hurtrez-Boussès S, Blondel J, Perret P, 1997. Relationship between intensity of blowfly infestation and reproductive success in a Corsican population of blue tits. J Avian Biol 28:267 270. Hurtrez-Boussès S, Renaud F, Blondel J, Perret P, Galan M-J, 2000. Effects of ectoparasites of young on parents behaviour in a Mediterranean population of blue tits. J Avian Biol 31:266 269. Johnson LS, Albrecht DJ, 1993. Effects of haematophagous ectoparasites on nestling house wrens, Troglodytes aedon: who pays the cost of parasitism? Oikos 66:255 262. Johnson LS, Eastman MD, Kermott LH, 1991. Effect of ectoparasitism by larvae of the blow fly Protocalliphora parorum (Diptera: Calliphoridae) on nestling house wrens, Troglodytes aedon. Can J Zool 69:1441 1446. Jones CJ, 1996. Immune responses to fleas, bugs and sucking lice. In: The immunology of host ectoparasitic arthropod relationships (Wikel SK, ed). Wallingford: CAB International; 150 174. Klaasen M, Lindström Å, Zijlstra R, 1997. Composition of fuel stores and digestive limitations to fuel deposition rate in the long-distance migratory thrush nightingale, Luscinia luscinia. Physiol Zool 70:125 133. Klasing KC, Johnstone BJ, 1991. Monokines in growth and development. Poult Sci 70:1781 1789. Klasing KC, Johnstone BJ, Benson BN, 1991. Implications of an immune response on growth and nutrient requirements of chicks. In: Recent advances in animal nutrition (Haresign W, Cole DJA, eds). Oxford: Butterworth-Heinemann; 135 146. Klasing KC, Laurin DE, Peng RK, Fry DM, 1987. Immunologically mediated growth depression in chicks: influence of feed intake, corticosterone and interleukin-1. J Nutr 117:1629 1637. Klasing KC, Leshchinsky TV, 1999. Functions, costs, and benefits of the immune system during development and growth. Proc Intl Ornithol Congr 22:2817 2835. Kyriazakis I, Tolkamp BJ, Hutchings MR, 1998. Towards a functional explanation for the occurrence of anorexia during parasitic infections. Anim Behav 56:265 274. Maier SF, Watkins LR, 1999. Bidirectional communication between the brain and the immune system: implications for behaviour. Anim Behav 57:741 751. Merino S, Moreno J, Potti J, León A de, Rodríguez R, 1998. Nest ectoparasites and maternal effort in pied flycatchers. Biol Conserv Fauna 102:200 205. Merino S, Potti J, 1996. Weather dependent effects of nest ectoparasites on their bird hosts. Ecography 19:107 113. Møller AP, 1990. Effects of parasitism by a haematophagous mite on reproduction in the barn swallow. Ecology 71:2345 2357. Møller AP, 1994. Parasite infestation and parental care in the barn swallow Hirundo rustica: a test of the resource-provisioning model of parasite-mediated sexual selection. Ethology 97:215 225. Møller AP, 1997. Parasitism and the evolution of host life history. In: Host-parasite evolution general principles and avian models (Clayton DH, Moore J, eds). Oxford: Oxford University Press; 105 127. Møller AP, Allander K, Dufva R, 1990. Fitness effects of parasites on passerine birds: a review. In: Population biology of passerine birds (Blondel J, Gosler A, Lebreton J-D, McCleery R, eds). Berlin: Spriner-Verlag; 269 280. Møller AP, de Lope F, Moreno J, González G, Pérez JJ, 1994. Ectoparasites and host energetics: house martin bugs and house martin nestlings. Oecologia 98:263 268. Munger JC, Karasov WH, 1989. Sublethal parasites and host energy budgets: tapeworm infection in white-footed mice. Ecology 70:904 921. Nelson WA, 1987. Other blood-sucking and myiasis-producing arthropods. In: Immune responses in parasitic infections: immunology, immunopathology, and immunoprophylaxis (Soulsby EJL, ed). Boca Raton, Florida: CRC Press; 175 209. Ots I, Kerimov AB, Ivankina EV, Ilyina TA, Horak P, 2001. Immune challenge affects basal metabolic activity in wintering great tits. Proc R Soc Lond B 268:1175 1181. Perrins C, 1979. British tits. London: Collins. Potti J, Moreno J, Merino S, Frías O, Rodríguez R, 1999. Environmental and genetic variation in the haematocrit of fledgling pied flycatchers Ficedula hypoleuca. Oecologia 120:1 8. Richner H, Oppliger A, Christe P, 1993. Effect of an ectoparasite on reproduction in great tits. J Anim Ecol 62:703 710. Richner H, Tripet F, 1999. Ectoparasitism and the trade-off between current and future reproduction. Oikos 86:535 538. Rothschild M, Clay T, 1952. Fleas, flukes and cuckoos. London: Collins. Saino N, Calza S, Møller AP, 1998. Effects of a dipteran ectoparasite on immune response and growth trade-offs in barn swallow, Hirundo rustica, nestlings. Oikos 81:217 228. Szép T, Møller AP, 1999. Cost of parasitism and host immune defense in the sand martin Riparia riparia: a role for parent-offspring conflict? Oecologia 119:9 15. Tinbergen JM, Boerlijst MC, 1990. Nestling weight and survival in individual great tits (Parus major). J Anim Ecol 59:1113 1127. Tripet F, Richner H, 1997. Host responses to ectoparasites: food compensation by parent blue tits. Oikos 78:557 561.